This week I interviewed Daniel Munro and Creig Lamb. Among their several hats, Daniel and Creig co-direct Shift Insights, where they work to “cut through complexity to provide clarity about the social, technological, and economic challenges and opportunities facing Canada”. In that role, they recently released “Northern Potential: Advancing Canada’s AI Ambitions”, a report that examines AI adoption in both the public and private sectors in Canada that is well worth a read.

Canada’s history with AI should be familiar to most readers: from leading the world in the research phase, to taking a backseat on bringing the technology to market, and now considering how best to adopt and use the technology in the context of a push for technological sovereignty after the fact. Daniel and Creig raise several key questions that Canada must answer as we work to adopt AI: asking what we hope to achieve, and why. They share insights into how Canada can begin to answer these questions as a guide to driving effective AI adoption in a way that is consistent with Canadian societal values, and make concrete suggestions through which the Canadian public sector can overcome practical challenges standing between Canada and the efficiency gains promised by AI. The values-based, human-centred approach they suggest in the report should be required reading for policy makers involved in any aspect of AI adoption.

While our conversation touches on several aspects of the report itself, we cover more ground, including issues of data sovereignty, intellectual property, and the “low-innovation equilibrium” in which Canada finds itself. The interview is not a substitute for the report, and I suggest reading both together to get the full benefit of their contribution.

Your email client will probably truncate this post. My key takeaways are presented at the end, so be sure to read the web version if you want to get the whole story. Many thanks to Sasha for taking the time to share his insights.

Interviewer’s note: Daniel Munro and Creig Lamb approved the final version of the section entitled “Interview with Daniel Munro and Creig Lamb” and had editorial input on that section, with the option to rephrase and expand on the ideas discussed in the interview without changing or removing any intended meaning. The key takeaways presented at the end are my own commentary, and do not necessarily represent the views of Daniel Munro, Creig Lamb, or Shift Insights.

Interview with Daniel Munro and Creig Lamb

KB: Tell us about yourselves and give the audience some context for where your perspectives are coming from.

DM: I do three main things. I’m co-director with Creig of Shift Insights, which is a policy research shop that examines the social, technological and economic challenges and opportunities facing Canada with a view to helping us become a more prosperous and just society. We look at a wide range of policy issues related to innovation, education and training, skills, and technology development and adoption.. It’s in that capacity that the two of us put together this report for the CSA Public Policy Centre. I’m also Director of Research and Innovation at Actua, which is a youth STEM outreach organization. We deliver STEM camps, clubs, and workshops to youth all across the country. And I’m a senior fellow in the Innovation Policy Lab at the Munk School of Global Affairs. I’m actually a political philosopher by training. The innovation and skills policy work came more recently, if 15 or 20 years ago could be called recent. That’s ancient history to someone Creig’s age. In any case, in the first part of my career I was a political philosopher and I spent a little bit of time in academia. I still keep the political philosophy hat in my back pocket and I keep one foot in academia thinking about innovation and technology ethics.

CL: My main policy hat is shared with Dan. I’m the co-director of Shift Insights and the chief economist. I come at it from an economics background primarily. I do a lot of the data analysis and all of those components. Before that I was at the Brookfield Institute which is now The Dais. That’s where Dan and I met and worked on a bunch of projects together. I was leading a lot of their tech innovation skills and automation research at the time. And in my spare time, I design lamps.

DM: That’s the humble way of putting it. The real way of putting it is that you’re a designer and entrepreneur focused on lighting solutions.

KB: Let’s start with a little bit of context on the report you published on AI adoption in Canadian public and private sectors. How did that come about, and what are the goals of the report?

DM: When the government appointed a minister for artificial intelligence and digital innovation and gave him the job of developing an AI strategy for Canada, Creig and I thought that given what we had done on innovation policy and especially technology adoption in both the public and private sector, it might be useful to distill what we had learned in the form of helpful guidance for the minister to develop the strategy. In talking with the CSA policy group, the scope expanded to offering advice for diffusion and adoption in both the public and private sectors, since our previous work has given us some insights into what does and doesn’t work in those areas. And so what was initially supposed to be something like a moderate-length letter to the minister with some helpful hints turned into a longer report that came out a few weeks ago.

Initially, we thought about releasing something shorter much more quickly, in part because it wasn’t going to be based on new research. It was going to be a curation of the kinds of things that we already knew that might be helpful. We actually anticipated a four to six week turnaround, but organizations move at the speed of organizations, and as we got into the work we realized something more robust would be more helpful, so it took us months instead of week and we produced a longer report. We’re grateful to the CSA group for supporting the work and shepherding the review, design, and publication with us. We think it turned out to be better than it would have been had we just thrown a letter up on our website.

CL: I would add that while the intention was for quick advice, we quickly realized that the pieces that we were summarizing and contextualizing are challenges that the government has been tackling for decades in some form or another. It remains relevant even in the new context of AI. I think even as different mandates and different policies come out, these will probably continue to be challenges that the government will be tackling in some way, shape, or form over the next decade.

DM: The core question that occupied our thoughts and organized the writing was that in light of difficult innovation performance in the private sector and a string of challenges with technology adoption in the public sector, what is the minister going to do differently such that we don’t get the same result that we’ve seen time and time again? AI has features that are different from previous technologies, but it’s similar enough as a technology, and in terms of the background conditions that drive or block adoption. We’ve seen this story before and we don’t want to repeat it. But we’ve got to do something different in order to get a new story. That was the orienting question for us.

KB: You suggest in the report that Canada has this window of opportunity to translate early leadership in AI research into sustained economic and social benefits. What do you think is a realistic niche in which Canada can lead in AI, given that we don’t have the capacity to compete head to head with the American hyperscalers and others in the space?

DM: Creig can agree or disagree with me, but I’ll start by saying that I share the pessimism. We’re not Pollyannaish about this by any means. We share the skepticism in that we’ve seen this story time and again. In this case we have moved about as slowly as we always move, so the window might be closing.

To zoom out a little bit, just to put it into context, when we’re thinking about AI in Canada, we need to make a couple of distinctions. We’re thinking about it in two ways. One is from a “tech-making” perspective, we might call it, and the other is a “tech-taking” perspective. Tech-making focuses on the technological innovation, commercialization, and sales side of things, and what can be done to spur better performance. The tech-taking side of things itself has two dimensions: adoption in the public sector for the sake of productivity and efficiency gains (as well as pursuit of key public sector values), and adoption in the private sector. So there’s actually three moving parts here.

We’re less pessimistic and less skeptical about the prospects for tech-taking and using technology, both in the public and private sector. We are probably a little bit more skeptical about the tech-making side of things. I do think that there might be some niche areas where we can offer things. One that really comes to mind is AI for supply chain management. Think about the Scale AI innovation cluster. We might ask whether we can develop ways of using AI to spot possibilities and challenges in global supply chains and then use that information to manage those supply chains better. We might also be able to export AI for supply chains solutions for use by other companies and countries. That’s an area where I think the window may be closing, but we might still have some opportunity to do things there - especially given ongoing disruption and changes in global supply chains.

I also think that there are probably opportunities to do niche things with a Canadian focus. Even if we don’t develop things that have global export potential, there are some things that we can do domestically around health technology. Given concerns about health data and privacy, “Canada first” solutions might be the way for health tech. You may not be able to export what you develop in that area, but there’s probably enough both technical know-how to do Canada-focused things, and political will that would be quite receptive to Canadian made solutions for health. There’s probably some defence related things as well.

CL: To bolster your point on the tech taking side of things, looking at it from a broad economic perspective, we’re really thinking about how to improve things like productivity with the goal of gaining the widest benefit in terms of wages, equity, distribution, etc. Canada has had a particularly deplorable track record on adopting technology writ large both in the private and public sectors, but there are some signals that AI with its ease of use and subscription-based models that require relatively low upfront investments may be able to break some of these trends. So we do think that there is an opportunity to facilitate the adoption of AI and promote diffusion in a more concerted way across the private sector that might look a bit different than some other forms of digital technology.

I think the opportunity, and where most of our effort should lie, is focusing on other sectors in which we have a competitive advantage or that are strategically important or that simply employ a lot of people and focus on making them more competitive through the use of AI. While I will always say that the development of technology in our homegrown tech sector is important we will never out compete the giants. There are places on the margin that we can grow, but I think really it is about helping the average firm recognize how they can compete more internationally and hire more people and just be more productive. Hopefully that wealth gets distributed in the form of wages and things like that.

KB: That leads into a tension that you explicitly call out in the report, which is the idea that the minister has a mandate to ensure Canada has the enabling infrastructure, data storage, compute, energy, etc., to use AI effectively. You acknowledge in the report and in the preceding discussion that we are probably going to be dependent on foreign hyperscalers for AI service provision, both in the public and private sectors. At the same time, there is a lot of talk about “technological sovereignty” as a guiding principle of policy development. What does sovereignty mean in this context, given our external dependencies?

DM: You’re asking a political philosopher to define sovereignty. Do you have two hours?

I’ll give you a really pithy answer. My understanding, or my way of thinking about sovereignty, centres on whether or not you can meaningfully shape your own future. When we’re thinking about meaningfully shaping our Canadian future, that doesn’t necessarily exclude all external influences or inputs. It just means that when you pull everything together, you need to be able to say that you had or have the power to make decisions about where you want to go. I think there’s a tendency to think that sovereignty requires absolute control over all components, products and uses, which is simply unrealistic given the reality of fragmented global value chains, how components are pulled together to produce products and services, and how no country can produce everything itself that it might require. But that’s not resignation to an uncomfortable reality. Relying on others for inputs can in many cases enhance our power to shape our own future, by allowing us not to get bogged down in things that we don’t do well. Ultimately, the core question should be: “can we meaningfully shape our own future in this space.”

That means that on some things, it may be okay to rely on the external hyperscalers. In other cases, it might not be okay to rely on external hyperscalers. It comes down to what kinds of data you’re using, where that data is stored, and the privacy requirements around it. It depends on whether someone can push you in a direction you don’t want to go because the control some vital component or input you can’t get elsewhere. In some areas, I’d lean more towards the “tech nationalist” approach – when it comes to health applications and health data – versus other areas where we’re simply crunching public data for economic reports or trends.

I think the second question that I would ask is simply “what is it that we want to achieve with AI or with any other technology”?

In the tech-taking space, we want to achieve efficiency, productivity, maybe some other goals like more accurate answers to questions that people have about government services. What are the tools that we need in order to do that, and can we maintain meaningful control over what the future of that activity looks like?

When it comes to tech-making, that’s where it gets a lot trickier. If we want to have homegrown firms that develop solutions both for domestic consumption and for global export, how do we maintain IP? What data do we draw on, and to what extent, if at all, can we use the products of the hyperscalers?

That’s how I would think about AI sovereignty, which is not an answer. It’s more like principles for how to think through what you want to do in any given case. I’m going to do that to you again on your other questions. I’m not going to give you an answer. I’m going to give you criteria.

CL: This is how we think about things: we come up with a framework and then send it out to the world. I would add, though, that I often wonder, from a public policy lens, how much we also impose our views, and views around things like tech nationalism, on individual firms, and whether they actually coincide with the views of the individual firms. Firms are probably mainly concerned with making a go at it and making a living and being competitive. The question of whose infrastructure they’re using probably matters less for the individual firm than the public sector writ large. So I wonder sometimes how much we impose those viewpoints on specific firms, and in doing so, to what degree do we put them at a competitive disadvantage in an industry in which they were already starting 10 steps behind.

That doesn’t answer anything, either, it just adds another question.

KB: I want to go back to a point raised by Daniel around intellectual property. Part of the minister’s mandate is to maintain broader societal values and democratic norms as we adopt AI. Large language models ingest massive amounts of data. They,, and for the most part ignore copyright to get it. On the other hand, if you’re going to build an AI system that has any kind of chance at competing with what’s out there already, you pretty much have to play that game. Is there a path to using LLMs that is consistent with tIP rights, or does this new technology require us to fundamentally rethink what copyright means?

DM: I’ll give you half an answer and tell you why I won’t give the other half answer. The half answer is that if we take an application or use-specific perspective on some things, I think there are ways to develop AI solutions that don’t necessarily violate copyright or expose intellectual property or personal data to risk.

I’ll make it more concrete. If you’re designing a chat bot for the Canada Revenue Agency that’s intended to answer questions about Canada Revenue Agency procedures, rules, and that kind of thing, you will be drawing on data that doesn’t require IP protection: CRA policies, procedures, and things like that feed into that. It could help answer questions that users might have about how to file their taxes or what exemptions would be permitted. So, on one hand, for certain kinds of applications, I think we can develop those without violating these other things. To be sure, the base model on which a CRA chatbot is built may have been trained on copyrighted data and thus have antecedent IP violations built in. That’s a problem. I’m not sure how you disentangle those things, to be honest. It’s certainly something that the government will need to keep in view.

I will say I lean more towards protecting copyright and intellectual property than I do towards changing rules simply to facilitate development of some new technology. That might be a personal bias. Creig is the second most creative and stylish person I know. The most creative and stylish person I know is my wife, who’s an artist. In the arts community, the attitude towards generative AI is deep skepticism. The pervasive view in the arts community is that AI is inextricably built on IP theft. And in those specific use-cases, I think that view is right. . I think both analytically, but also for personal reasons, I lean more towards the idea that, if there is a tension between these two things, between protecting IP and developing AI, we should protectIP rather than allow violations to facilitate the development of AI services and products. That might suck for AI developers, but we all have to innovate under constraints.

CL: I concur with that. I think this report came at it from the lens that what already exists is a given, that we can’t dial it back, and then considering what we can do in the current environment in terms of adding additional measures and protections. That’s the subject for a whole other report that I would be willing to dive into.

KB: You gave a few examples of applications, and the mandate given to the minister is to push towards efficiency and effectiveness in the public sector. On the other hand, LLMs match patterns and produce likely text, not necessarily correct text. How do you balance the drive for efficiency against the reality that AI systems hallucinate and have built-in biases? Is it possible, in the framework suggested by the guiding principles that you develop in the report, to identify the kinds of tasks that can be handed off to AI versus the stuff that should remain in human hands?

DM: This is the other question where I’m to give you two criteria rather than an answer.

In the public sector, when we’re thinking about whether or not to adopt AI to make some government function or service more efficient, yes, there’s the concern about whether or not the technology will hallucinate, and produce inaccurate answers with an enormous amount of confidence, which is actually the real problem. It’s not just that the outputs are wrong sometimes, it’s that they’re wrong and there’s no function for admitting they’re wrong. So one principle or a question that I would ask is whether, for any given task, the use of AI is more or less accurate than the current way of doing that task. Again, take the CRA example. Is a CRA chatbot any more or less accurate than getting a CRA agent on the phone? Because we know they’re not perfect. Let’s say that 4 or 5% of answers from a chatbot are inaccurate. If it turns out that this is more accurate than the people who were doing that job, we should be cautiously OK with that, and deal with reality rather than perfection, given that existing systems are also imperfect.

There is one caveat, before I give you another principle. Human-centered approaches are better at learning from error and improving than chatbots. If you tell a human agent that they got something wrong, they have a capacity to integrate the better answer into the mix pretty quickly. By contrast, getting a chatbot to acknowledge and learn from errors is a much more involved process. And if it produces a better answer later, it’s still not doing that because it has selected the right answer as the right answer, it selects the right answer because it’s the most likely answer, or it’s been given a specific instruction to produce a specific response even if its underlying model doesn’t recognize it as the most likely response. In short, I think the human capacity for learning from error is still superior to that of chatbots.

The second principle for thinking about this is to reject the idea that one is replacing the other and think about whether there are ways that the technology can complement the existing human approach to collectively produce a better outcome. To beat this example into the ground, maybe what you get is a CRA agent who answers a question, but before actually giving it, checks it against what the chatbot would produce. If the answers align, then your likelihood of accuracy is probably much higher. If they don’t align, then the CRA agent has to either think about whether they’re wrong or whether the chatbot is.

The funny part about that is that it would be a lot less efficient, but it might be more accurate. So this is one of these cases where you have to step right back and ask what it is that we want to achieve through the adoption of technology. We often say we want better efficiency, but we might also want to improve public sector accuracy. Those might not always go hand in hand.

CL: When you were saying in terms of supplementing existing human tasks and increasing efficiency, we’re really talking about fundamental automation, and that is how we’ve gained efficiencies and productivity growth throughout the entirety of technology adoption throughout all of history. There’s a famous example of the ATM and the bank teller where everyone thought the ATM was going to replace the bank teller completely, but all it really did was remove the transactional, basic, routine-oriented handing of cash as a job task for the bank teller. Their job became actually more complex and value added in terms of customer relations and financial advice and things like that. Automation allows people to remove some of the routine things and do tasks above and beyond what technology can do, which overall leads to efficiency gains. The more we can see people eliminating some of the routine tasks that would take them all day so that they can do more complex things that day, the more we will see efficiency gains throughout the government.

I think it would be a useful exercise for people to go through within government and look for areas where the most efficiencies can potentially be gained with like an eye to some quick wins to gain momentum with the least amount of potential risk.

I think we cannot separate the skills question from this area. As we’ve learned with AI, and the more I talk to folks who are using AI and employing AI within their companies, the more I realize it actually takes quite a bit of labor to verify the output and check and ensure that hallucination hasn’t occurred and actually understand the mechanisms, to some degree, of how AI is working, how to ask it questions, and how to query it properly. I think there’s going to have to be a lot of effort put toward building that skillset within the government. That’s been a shortcoming of the government for a long time.

DM: I’ll add one thing to not the last thing that Creig said, but the one previous, riffing off the example of introducing ATMs and how that actually allowed tellers to move up the value chain. (James Besson, Learning by Doing). I 100% agree that when done well, you introduce a technology that can automate the lower value tasks and thereby liberate employees to do higher value tasks. This is good for them, good for the business.

Here’s where I’m going to get myself in trouble: I’m not convinced that that’s the perspective or orientation that the government is taking around AI. What we’re seeing is an AI strategy at the same time that we’re seeing an attempt to reduce the size of the public sector. There’s this hyper focus on the number of public servants we have, or the ratio of public sector employees to population. And there’s a sense among many people that the number is too high or the ratio is out of whack, so we need to fire a bunch of civil servants. My sense is that the government isn’t thinking in the sort of James Besson “ATM-moving-employees-up-the-value chain” way. They’re thinking about how AI can fill the roles of public sector workers that they want to take off the payroll, and that concerns me.

As we note in the report, the public service has many values that it must uphold. These include respect for people, respect for democracy, integrity, stewardship, excellence, equity and accessibility. Reducing the size of the public sector by adopting AI might get you some efficient gains or cost savings. Maybe. Big maybe. We also need to think about the impact on these core values. For what it’s worth, upholding these values likely requires a lot of inefficiency and redundancy. The public sector isn’t the private sector and we shouldn’t treat technology adoption in these sectors as equivalent processes.

CL: The automation side of things is kind of where I got my start in public policy and innovation research. I took a historical perspective on this, and throughout history, the most successful and long lasting versions of automation have been labor-augmenting, not labor-replacing. I think looking at technology as augmenting human capacity rather than replacing it is actually the smartest way to do it, and clearly the most ethical, and that’s the most important piece, but even if you’re looking at it from the perspective of pure efficiency gain, you should think about augmenting tasks, not replacing entire jobs.

KB: Let’s dive into the skills question. Clearly there is a difference between the stated intention and the approach that’s being taken to public sector adoption of AI. Is this the result of missing skills in the government to properly utilize this technology? What gaps do you see?

DM: I think Creig can probably speak to this better, but I’ll take a swing at it anyway. With respect to using AI in the public sector, there’s obviously some skills required in order to do that. In the report, we lay out three categories of skills that need to be considered.

One category is technical skills: you need people who have the technical and technological skills to develop, implement, and maintain AI solutions.

Then you need user skills. To go back to our CRA example, the CRA agent might not have to be an expert in how the chat bot actually works, but needs to be able to make the right queries, offer the right prompts, and have the capacity to critically evaluate whatever comes out the other end.

Finally, you need management skills: the skills to properly manage technology transformations, which is partly, if not largely, what a lot of this is about. Management skills to procure solutions, to manage the transformations, to ensure that employees have the other categories of skills, and things like that. So you need technical skills, user skills, and management skills to do all of this right.

One of the things that we’re concerned about is that at the same time that the government has shortages, especially in the technical skills side of things and to some extent on the management side, they’re actually cutting the public service. The Treasury Board says 30% of digital roles in the government are currently vacant. There’s already a technical skills gap. At the same time that they have those vacancies, they’re letting people go. What I’m concerned about is the signal that sends to people who might fill those roles that jobs in the public sector are precarious. The government already faces challenges around whether they can pay the same rates to technical people as the private sector. If you’re on the technical side of things, you might have good software skills, and you’re thinking about where to go. If the government says they need some technical people, but you’re also reading that the government is just getting rid of people, seemingly left, right, and center, why would you go work for the government? My concern is that they’re almost shooting themselves in the foot here, where stated need does not align with the signals they are sending.

CL: I think that’s a good point. You’ve seen it time and time again. As the government increases its focus on sort of digital technical talent they’re increasingly competing with the private sector. What’s the value proposition for people with either highly technical digital skills or just general digital literacy, who could work in all sorts of private sector roles that would pay more, probably have more flexible hours, all sorts of benefits, maybe even in cities that they want to live in or remote. So why would they work for the government? Part of the answer is job security, but if they’re sending the signals around a lack of public sector job security or a shrinking workforce, that undermines the value of that component.

I think the management side is another really big piece, and building up skills around procurement and oversight for digital transformations. We’ve seen so many challenges with respect to that. How can they improve their capacity and skills in areas that enable them to both identify areas where digital transformation should happen and then manage the procurement and development and diffusion of the results.

KB: Your report discusses how Canadian firms have been stuck in a “low innovation equilibrium”. Can you elaborate on what that means? Given that this issue predates the recent AI explosion, has the situation been changed at all, or could it be changed, by AI?

DM: My first experience with the phrase “low innovation equilibrium” came from a paper written by Peter Nicholson in which he tries to explain why business innovation is so low in Canada. His answer to that question is that businesses are only as innovative as they need to be, because, he says that a lot of Canadian businesses have been able to maintain reasonable profits without having to innovate. Canada has the conditions, whether it’s market conditions, the cost of labour, or what have you, in which they’ve been able to maintain reasonable levels of profit without innovating. If you can maintain profits without innovating, why rock the boat? Firms simply don’t have any incentives to innovate.

Creig and I take this a step further. We have a conceptual model that breaks a couple of these things apart. We agree with the Nicholson line of thinking, but when we think about why firms do or don’t innovate, one side of the equation is whether or not they have incentives to do so, and the incentives in Canada tend to be weak. There is that reasonable profitability. There’s weak competition. If you’re not facing a lot of competition, why would you innovate? So there’s weak incentives.

The other side of the equation is limited capacity. Even if you’re a firm that wants to innovate you might lack the resources: financing, expertise, etc. In the past, a lot of government efforts to spur innovation have focused only on that limited capacity issue, for example by offering financing or tax credits or expertise around entrepreneurship and digital technology adoption or whatever. I think this is changing, but I think in a lot of cases, government policy doesn’t spend a whole lot of time thinking about the incentive side, or at least if they think about it, they haven’t taken any sort of strong measures to address it.

There is some noise around competition and improving the competitive environment, but the actual steps to do that haven’t really been as robust as I think they need to be. My worry is that to the extent that AI might contribute to innovation and productivity , a lot of firms are facing these weak incentives and in some cases, limited capacity to actually do things. If you want to spur adoption and innovation in AI, you have to begin thinking about improving those incentives before addressing the limited capacity issues. Focus on the drivers before we focus on the supports. We haven’t done that in the past, and I think we’re making that mistake again with AI. We’ve used the same conceptual framework for a number of things that we’ve done.

CL: While their AI has a particularly strong hype around it, the available data on adoption would suggest over the past few years that AI isn’t much different in terms of incentives and capacity challenges. We’re facing similar issues in terms of adoption. Adoption is relatively low and dispersed across the select few industries and large firms, those with the incentives and capacity to do so, so it doesn’t feel particularly different than other versions of digital technology.

DM: With one caveat though. When you look at the Canadian survey of business conditions, across the economy, 12% of firms say that they have adopted or will be adopting AI over the past 12 or next 12 months (which is up from 6% so it’s accelerating, but it’s still just 12%). This is what you get when you ask people who are authorized to speak about what firms are doing. When you talk to employees, though, the numbers are over 50%. Officially, 12% of firms might have plans and strategies and be thinking about this intentionally, but meanwhile something like half of employees in the economy are using AI, whether their firms are on board or not. This “shadow AI use” is a source of risk, obviously, so there’s actually a funny thing happening with AI, or at least certain kinds of AI, where firms are as conservative as they have always been with every other digital technology, but employees seem to be racing ahead in ways that we haven’t seen with previous technologies.

CL: This is a really good point. We’ve also seen a corresponding thing that does change the incentive structure. Companies are realizing that this is happening and are deciding to permit the use rather than developing a more deliberate strategy. The incentive is to let adoption happen by itself and just pay for the subscription or whatever else needs to happen to make it possible. That’s an interesting point of differentiation from previous technologies.

KB: In the report, you allude to fragmentation in the approach to public sector AI adoption: you identify vertically siloed departments, thousands of legacy systems, and poor application health, and you suggest a centralized body as a means to deal with it. What drives that fragmentation, and what does the centralized approach that you suggest look like practically?

CL: I think what drives the fragmentation is just the nature of the bureaucracy and the risk aversion and lack of communication and all of these sorts of bureaucratic limitations and rules. I have personally noted, on projects where I was working with two departments, that often government departments that are working on very similar things don’t know anything about the efforts of the other.

I think another piece is just the size and the mandates of these departments and the number of things that they’re doing. It’s really difficult for them to continue to keep on top of what everyone else is doing and coordinate and consolidate efforts.

There also tends to be a tendency to reinvent and build the same thing over and over again, as opposed to checking who’s doing what and partnering.

DM: For context, I think the fragmentation in the public sector predates any technology adoption. I mean, that’s just a feature of bureaucracies that are oriented to different functions and given different lines of power. So the fragmentation predates AI. The fragmentation or differences in technology adoption and systems and things like that is just a reflection, a symptom, of that deeper fragmentation.

The upshot is that what you get is six different departments that are trying to do the same thing end up with six different ways of doing that same thing because they don’t talk and because they have their own unique preferences about what exactly it is that they want a technology to do or a solution to be, and so you end up with these thousands of legacy systems that are different across the different departments.

An AI strategy isn’t going to solve the underlying problem of public sector fragmentation. Our hope is that if you can create or establish some kind of central body that will support and incentivize communication across departments, across ministries on uses of technology that appear to be similar, that you might get something that is a little more consistent across the different parts of the public sector. There is a suggestion about central coordination, information sharing, shared systems, and procurement of solutions that can be used by different parties in the report where AI is concerned. There’s a hope that that kind of thing can be done, but we’re adults about this. We know that this isn’t going to address the underlying problem. That’s a much, much bigger problem.

CL: You’re right. I do recognize that in a lot of instances, funding mechanisms for departments drive that kind of siloed approach. I also understand that there are different privacy issues across departments that require these sorts of siloed and fragmented legacy tools and systems, or at least explain this kind of thing. I think it is the hope that a central body could cut through a bit of that noise and identify common areas and common issues across departments and identify certain areas where it makes sense to adopt AI and ways to mitigate the risks that are common across departments, allowing insights from one department to inform another.

The issue is that we’ve also seen this happen and seen this done a number of times with central bodies and it often doesn’t play out in the way we think it will. Sometimes we end up with something that lacks the teeth it might need to be effective, and so it ends up being ignored.

KB: It strikes me that something that has an overarching awareness and oversight of all of the various is an ideal use case for an AI system. I’m curious if that’s something you thought about: is there a path to having an AI system actually do the convening work, or would we just be creating an Ouroboros of fragmentation?

CL: I don’t know why I didn’t think of that. It would be great for the central body to use AI.

DM: I think AI might be able to help with this, but having a well-staffed central body with the right mandate and authority to develop policies, share ideas and best practices, coordinate procurement and implementation, is the key piece. I can think of a couple of things that are already in place that suggest some precedent or pathways for a central body. We’ve got the Canada Digital Service. They should be a candidate for being part of something like this. But they’re wildly understaffed. They will do a project here and there and they can’t keep track of the whole. If you overlay some kind of function that has that sort of broader perspective across the whole public sector and not just individual projects that might be helpful.

You’ve also got the Treasury Board, which has developed the policies on automated and algorithmic decision making, so it’s not like it’s an idea that comes out of nowhere. There are things that could be transformed into this role. Maybe AI becomes part of this as well to help with fragmentation – but I think we need the agent and authority before we bring in the technology to help.

KB: Your report talks about Innovative Solutions Canada and compares it to the SBIR program in the US. This is a comparison I’ve made in other contexts. What creates the divide between the technological driver that SBIR has been for the US for the last 40 odd years versus the relatively lower impact that ISC has had in the Canadian ecosystem?

CL: Kyle, I know you’ve written about this. I like your take, so I’m going to mirror back some of your takes to you. I have a few things that I will add to what you’ve said.

The Innovation Solutions Canada was obviously modeled after the SBIR and there have been recommendations to adopt an SBIR-type of program floating around, both provincially and federally, for a long time. SBIR is one of the mainstays of successful innovation policies out there, along with DARPA and others.

You can see that they replicated some of the foundational principles with Innovation Solutions Canada. 21 departments are mandated to spend 1 % on procurement and R&D. As you’ve seen, these departments were consistently underspending by two thirds of their actual mandate. In 2024 the budget got slashed to match what departments were actually spending to about a third of what it was before. ISC didn’t really have the teeth of the SBIR, which is another kind of pre-commercial procurement where departments across the U.S. that spend $100 million or more on R&D have to spend roughly 3% of their R &D budget on small and medium sized enterprises.

I think there are three explanations I have for this. One is just fundamental risk intolerance. You can just see in governments across Canada that there’s just a lower tolerance for the kind of risks that comes with pre-commercial procurement and working with small innovators. You can see that in the kinds of companies that the Canadian government works with and our preferred vendor lists. They really skew toward major corporations that handlea wide array of government activities and can provide a turnkey solution. Our ability to work with one-offs and tolerate project failures is fairly low. I think there’s a bit of a cultural thing and a bit of a government culture thing that would need to be addressed somehow.

There’s also the issue of governance of the SBIR versus the ISC. Why would a program that’s being mandated to spend a certain amount not actually achieve its mandated spend. If you’re being forced to do something, why would you not do it? Why can’t it be enforced? I think this comes down to both a governance issue and a capacity issue.

Canadian government departments just have very low capacity for even basic things like moving away from “waterfall” procurement to more more agile procurement. So working on pre-commercial procurement with smaller firms on risky projects is not within the realm of knowledge and expertise of most departments. The string of big failures on the procurement side of things is kind of evidence to this.

It’s not like the US doesn’t have this issue, but you can see a slightly different governance structure: ISC is a central body that’s mandating government departments. SBIR has that central body, but then it has 12 other bodies within individual departments that manage and facilitate the actual guidance and oversight of the actual day-to-day of the SBIR implementation within the department. The DOD, the Department of Health, etc each have their own SBIR bodies that help facilitate this. I think that helps these departments actually work with smaller firms and commit to spending on the pre-commercial procurement.

You would have to not just mandate. You have to have these disperse networks of people with expertise embedded within the organization working on a day-to-day basis with the people that are managing their programs and the projects and the individual procurement to actually reach the numbers that you want.

And finally, as you’ve pointed out, there is the issue of scale. SBIR funds more than 5,000 projects in phase one and phase two. ISC funded something like 30 in the same timeframe. ISC is 1% of R&D spend and SBIR is 3.2%.

Altogether, ISC is a sort of “uncanny valley” type of program where on the surface, it looks like it was the SBIR, but as you start to dig, you realize it is not, and the differences turn out to be really important.

KB: Is there anything I should have asked you but did not?

DM: The only thing that I’d add or emphasize, that we say in the report, is that we’d really like to see not a technology first strategy on AI, but a values first approach. Rather than taking this technology and trying to find places where it might add value or advance something that we care about, instead ask “what is it that we want to achieve, and to what extent and how can this technology help us do that?”

Whenever people ask my friend and colleague, Danny Breznitz, what Canada’s innovation policy or strategy should be, he always says that the first thing we need to do is answer the question “what do we want from innovation?” Only then can we talk about an innovation strategy. I think that applies here as well. What is it that we want? What is it that we want to achieve?

We’re going to disagree about that, obviously. We have different values, we have different priorities. But once we come to some kind of answer or conclusion about that in the public sector and in the private sector, only then can we figure out what role AI could play in getting there.

A values-first rather than a technology-first approach to this is the way to go.

CL: That’s a really good point. Even on the low innovation equilibrium piece, I think sometimes we assume that it’s a problem because people think that innovation is intrinsically good. They want to bolster their R&D numbers so that they look good on an international stage or something. But if the end goal is to have people earning incomes and companies earning profits to facilitate those incomes, if that’s happening with or without innovation? Does it really matter?

DM: Innovation is just a mechanism. It’s an instrumental good. We have to ask the more fundamental question about what it is that we want to achieve, and then figure out how innovation fits into that.

Key Takeaways

The report by Shift Insights, and the commentary in the interview above, paints a complex picture of AI adoption in Canada. Aside from the usual challenges of fragmentation and siloed attempts to move forward that characterize pretty much all of Canada’s attempts to innovate, the core of what I took away from our conversation was a need to take a step back and answer a few fundamental questions before trying to do anything else.

What are we trying to achieve?

Daniel describes himself as a political philosopher, while Creig is an economist. It’s a powerful combination: throughout their report and this interview, they use a values-based framework to inform their recommendations, grounded in the realities of the incentives that dictate behaviour. Instead of prescribing a particular solution to Canada’s AI adoption challenges, their approach is to ask thoughtful questions that help us understand why we care about AI adoption in the first place and how the incentives might be misaligned, and only then to make concrete recommendations as to what needs to be done. It is an approach that I deeply appreciate, and use in my own work anywhere that systemic change is part of the path forward.

While this approach shows up throughout both the interview and their report, it appears most explicitly in our discussion of sovereignty. There is a huge amount of commentary making its way through the innovation space on what it means to have technological sovereignty and data sovereignty, and while these concepts seem straightforward on the surface, they are fairly difficult to define in practice. Daniel gets to the heart of what sovereignty means in the context of AI in just two questions:

“can we meaningfully shape our own future in this space?”

and

“what is it that we want to achieve with AI, or with any other technology?”

In other words: part of the reason for the difficulties Canada faces in its AI adoption journey is because we have not agreed on what is important. Daniel makes clear that he does not necessarily have answers for these questions to suggest; rather, that these are “principles for how to think through what you want to do in any given case.”

This idea generalizes far beyond AI, a lesson that I learned more slowly than I care to admit. Early in building my own company, mentors and advisors tried hard to educate me on the importance of clearly articulating “mission, vision, and values”. As a physicist who had not yet made the transition to business thinking, I mostly ignored their advice as tech-bro nonsense. In the end, it turns out I was wrong, and the tech-bros were right (broken clocks, etc.). Daniel and Creig’s questions invite us to articulate a vision (what we want to achieve long term), a mission (how we plan to get there), and values (the principles we agree are important to uphold in making decisions en route). Only once we have agreed on what these are does it make sense to decide how to go about adopting AI, or building Canadian data sovereignty, or, frankly, almost anything that requires significant change.

It is a framework that is particularly effective for complex decision-making, since values, in particular, provide guidelines that allow us to make decisions with ambiguous implications, secure in the knowledge that the decision was made with purpose and that we have guardrails with which to course-correct when things get off track enroute.

Shaping our future

I am personally skeptical of Canada’s ability to lead in the AI space, and have expressed as much in contributions to the innovation discourse. As far as I can tell, AI is just the latest in an often-repeated story in Canada: we contribute foundational, groundbreaking research to the world, and then sit back and watch as someone else makes it practically useful.

That being said, Daniel and Creig make a strong case that there are still opportunities for AI to create value for Canada and to direct our own future in the space, even if those opportunities may not involve setting the global agenda. They see models that are purpose-built for Canada as being an important area of focus, built using proprietary datasets that are specific to the society they serve. Health data is a key example, as is anything that similarly depends on sensitive, personally identifying data for an associated AI model to be useful:

“[T]here are probably opportunities to do niche things with a Canadian focus. Even if we don’t develop things that have global export potential, there are some things that we can do domestically […]. Given concerns about health data and privacy, “Canada first” solutions might be the way [….]”

The suggestion from the interview is to focus on models trained on Canadian data, geared toward solving domestic challenges.

While the suggestion is sound, it reinforces the importance of addressing issues of control over data that have plagued the space to date. Our ability to shape our future in AI hinges on being able to control the data on which that AI is trained.

Challenges around data sovereigty go beyond personal information, however. As I have previously contributed to writing, our ability to shape our future depends on control over our intellectual property, and AI does not have a great track record of respecting it. A key issue that arises when considering Canadian specific data is the need to fine-tune existing models rather than training them from scratch, given the volume of data involved. Even if the data that we use to train a Canadian-specific model is managed in a way consistent with our values, the base models almost certainly have not been. Per the interview:

“the base model […] may have been trained on copyrighted data and thus have antecedent IP violations built in. That’s a problem. I’m not sure how you disentangle those things, to be honest. It’s certainly something that the government will need to keep in view.

I will say I lean more towards protecting copyright and intellectual property than I do towards changing rules simply to facilitate development of some new technology.

Despite this preference, Creig takes a pragmatic view:

“[T]his report came at it from the lens that what already exists is a given, that we can’t dial it back, and then considering what we can do in the current environment in terms of adding additional measures and protections.”

Is productivity the end goal?

Canada is stuck in a “low-innovation equilibrium”, an idea that has been around for some time. Unlike most commentators on the issue, however, Daniel and Creig do not take for granted that an innovative Canadian economy and the associated productivity are the end goals, asking that we go one step further and first ground ourselves by articulating why this matters. As Creig puts it:

“I think sometimes we assume that [low levels of private innovation is] a problem because people think that innovation is intrinsically good. […] But if the end goal is to have people earning incomes and companies earning profits to facilitate those incomes, if that’s happening with or without innovation, does it really matter?”

Daniel echoes the point:

“Innovation is just a mechanism. It’s an instrumental good. We have to ask the more fundamental question about what it is that we want to achieve, and then figure out how innovation fits into that.”

Especially where the public sector is concerned, the answer to these questions has practical consequences for what it means to adopt AI at all. Daniel and Creig observe that we need to go through an additional layer of reflection on what we want, since there are tradeoffs to be made even after we have decided to proceed with AI:

“[Suppose a CRA agent] answers a question, but before actually giving it, checks it against what the chatbot would produce. If the answers align, then your likelihood of accuracy is probably much higher. If they don’t align, then the CRA agent has to either think about whether they’re wrong or whether the chatbot is. The funny part about that is that it would be a lot less efficient, but it might be more accurate. So this is one of these cases where you have to step right back and ask what it is that we want to achieve through the adoption of technology. We often say we want better efficiency, but we might also want to improve public sector accuracy. Those might not always go hand in hand.”

In other words: efficiency may come at the cost of accuracy, and vice versa. There is not necessarily a “right” balance, and even if there is, it will be highly context-dependent. A clear goal and values to guide us toward it are required to arrive at an answer that will serve Canadians, and getting these right has real consequences:

“There’s this hyper focus on the number of public servants we have, or the ratio of public sector employees to population. And there’s a sense among many people that the number is too high or the ratio is out of whack, so we need to fire a bunch of civil servants. […] They’re thinking about how AI can fill the roles of public sector workers that they want to take off the payroll, and that concerns me.

Only when we come to a consensus on how we want AI to benefit Canadians will we be able to move forward with AI adoptions, grounded in the values we have collectively decided are important. I do not think that blindly chasing productivity for productivity’s sake is the answer. The report by Shift Insights makes several recommendations on how to approach the necessary questions that should be required reading for anyone in the public service that is involved in AI adoption efforts.

Many thanks to Daniel and Creig for taking the time to share their insights, and for their contribution to the national conversation.

This week I interviewed Sasha van Katwyk, Senior Health Economist and Managing Principal at the Institute of Health Economics (IHE). As Sasha describes it, “IHE is a not-for-profit research institute that […] sits at the intersection of academia, government, and industry”, providing insight to policy makers and innovators alike on the value and potential of new health technologies.

This interview is the second instalment in a series on “purposeful research”, focused on finding ways to identify research projects that address real societal needs. My goal in this interview was to learn how health innovation is valued and evaluated by the different stakeholder groups in the adoption process, and to better understand what tools can be applied early in the innovation process to focus healthcare R&D investment where it can achieve the greatest impact.

Sasha’s unique perspective at the intersection of academic research, private sector health innovation, and healthcare policy will be of interest to anyone operating anywhere in that continuum.

Your email client will probably truncate this post. My key takeaways are presented at the end, so be sure to read the web version if you want to get the whole story. Many thanks to Sasha for taking the time to share his insights.

Interviewer’s note: Sasha van Katwyk approved the final version of the section entitled “Interview with Sasha van Katwyk” and had editorial input on that section, with the option to rephrase and expand on the ideas discussed in the interview without changing or removing any intended meaning. The key takeaways presented at the end are my own commentary, and do not necessarily represent the views of Sasha van Katwyk or the Institute of Health Economics.

Interview with Sasha van Katwyk

KB: Tell us about yourself and give readers some context for where your perspectives are coming from.

SVK: I’m an economist by training, specializing in risk assessment to inform social protection policy. I use simulation modelling to allow for the combination of disparate evidence sources to inform economic decision-making and forecast outcomes. I first used that skill set while working in South Africa to do simulation models for cash transfer programs; think welfare payments, anti-poverty initiatives, and taxation policy, things like that. When I moved back to Canada, I got into health research because there’s a growing area of simulation modelling in health economics. That’s where I first got involved in health policy, about 10 years ago. I worked within a research institute that was academically oriented, doing in-hospital trials and hospital-level policy-making. I then moved to the Institute of Health Economics (IHE) where I’m working on policy analysis and research for provinces, innovation investment centres, and formulary committees, all asking questions about health system capacity and decision analysis.

KB: What does IHE do, and how do you fit into it?

SVK: IHE is a not-for-profit research institute that we often say sits at the intersection of academia, government, and industry because those are the three sets of collaborators that we usually work with. Our work is a mix of grant-funded research and service-specific consulting. We also provide independent third-party analysis and advice for health policy and investment choices.

In my role as Managing Principal, I’m involved in a couple work portfolios; Health Systems Modelling, where our clients are either provinces or specific health condition foundations, Innovation Support, where we work with SMEs and independent researchers to define their value proposition and identify the best evidence generation strategies to build their economic case for system-level adoption, and Methods Advancement, where we just nerd out on the latest mathematical modelling methods and service delivery strategies in order to keep the organization at the forefront of quality analytics and delivery.

KB: Sitting at the intersection of those three stakeholder groups, what kind of disconnects and misunderstandings do you see in terms of how they are approaching health policy?

SVK: Everybody has their blind spots, which lead to misunderstandings and incorrect assumptions about what others’ priorities and needs are.

With researchers, the main blind spot is that they don’t understand the set of constraints under which policymakers operate. Researchers tend to present problems and solutions in a way that’s very “blue sky” policy-oriented, that imagines system-wide change. Even when the changes seem relatively modest to them, they’re often not aware that implementing what they are proposing—what they identify as the optimal solution—would require coordination and collaboration across many agencies within a government body and would consume an enormous amount of political capital to achieve. Because researchers don’t fully appreciate the policy environment they are working within, and the political and operational pressures within government agencies, there’s a disconnect between how they frame problems, how they present potential solutions, and what is actually practical from a policymaker’s perspective.

From the policymaker’s perspective, in my experience, they often don’t have access to the operational or technical details of the problem and aren’t operating with a deep understanding of the environment they are regulating. They don’t know these problems in depth. Often this is because, at the level where decision-making actually happens, policymakers think in terms of the policy levers at their disposal. If a problem can’t be addressed by pulling one or some combination of those levers, it’s very hard for change to happen without legislation, and now you’re demanding a lot of the agency or department you’re engaging with. As a result, there is a strong institutional gravity toward existing options. If the problem is not solvable through a relatively straightforward policy lever, it is difficult to sustain policymaker engagement.

For innovators, especially independent or small-scale groups, I think the main blind spot is that they don’t understand the decision-making process and what policymakers are using to assess the value and viability of a technology that they might want implemented. In part that’s not their fault, but it’s still a significant blind spot. Innovators often envision that the system works one way, but in reality it operates very differently. As a result, innovators are often unaware of what actually drives meaningful change at the system level.

KB: In IHE’s role at the intersection of all those blind spots, what have you found to be effective at convening all of those viewpoints and getting them onto the same page?

SVK: We are most successful when there is alignment with policymakers who know they want to resolve a problem and come to us with an understanding of the levers at their disposal, and are looking for solutions that fit within those levers. Often, we have to get them to that place before any math can actually happen, and it can be where the real intense work has to happen. Once we have this clarity, it is easier for us to then look at the research, do the deep dive, develop a nuanced understanding of the nature of the problem, bring in the researcher perspective on the true nature of the problem and the available solutions, and assess where those solutions fit within what is legislatively possible, what is practical, and what is financially feasible at the policy level.

We’ve definitely produced analyses that are missing one of the major actors, and we’re trying to speak to that actor as part of the process. That can be useful, but it’s much more challenging if there isn’t already a high level of clarity about what is possible, who is already invested, and who is willing to commit capital to the effort. When you don’t have that buy-in to actually implement change, it makes the whole process significantly more difficult.

KB: Walk me through the decision-making process that occurs when a policymaker is evaluating a novel health technology. Where are these policymakers sourcing problems, and once a possible solution is identified, what does the evaluation process look like?

SVK: That’s a really big question. There are a couple of caveats that need to be made as part of answering it. One is that not all health technologies are pitched as a solution to a clearly defined problem. It is often the case that there are technologies that have industry champions or policy-oriented champions who have identified a specific health technology and have the necessary network within government to drive a decision forward somewhat independent of a formal problem-identification process. That absolutely happens. So an important caveat is that when policymakers already have a solution in mind, the process is very different than when they are recognizing a problem, or appreciate that a problem exists but don’t necessarily know what the solution is or what technologies could be implemented. It can also be the case that even though a particular problem has not been formally identified as a policy priority, an innovator approaches the government with a health technology solution. You’re running into very different versions of the government decision-making process in each of those cases.

The other caveat is that when we say “health technology,” we need to be more specific. There are different processes for different kinds of health technologies. If we’re talking about pharmaceuticals, for instance, there’s a relatively well-established federal process for regulatory approval and health technology assessment. That process then filters down to the provincial level, which is where coverage and reimbursement decisions are actually made. Broadly speaking, it involves a national-level process of going through Health Canada for regulatory approval, then the Canadian Drug Agency for clinical and health technology assessment, and then the pan-Canadian Pharmaceutical Alliance (pCPA) for price negotiation, and only at that stage do provinces begin making reimbursement decisions.

There are other technologies—for example, digital technologies or medical devices—that might or might not go through Health Canada, depending on how they are classified. I’m not going to pretend to be an expert in how different health technologies are classified, because it can be a gray zone in some cases.

Another unfortunate complication in answering your question is that the means by which decisions are made about whether and how to adopt a health technology are often quite ad hoc. Every province has its own process, and the level of development and maturity of that process is highly variable. In some cases, the decision is not even made at the provincial level; instead, it may be made at the health system or regional health authority level, with organizations making adoption decisions independently. Even when it is a provincial decision, the kinds of evidence being presented and what the pathway toward adoption looks like can vary substantially, and sometimes the process is not especially well defined, even within a given province or territory. It might not even be clear to policymakers themselves what the exact process is in a given case.

Basically, every part of your question has a number of asterisks attached to it. It depends what you mean by health technology, it depends what you mean by decision-maker, and it depends what you mean by adoption.

KB: Let’s simplify by using concrete examples of the kinds of analysis that IHE does — walk me through the inputs and outputs of your simulation models, and how they help inform decisions downstream.

SVK: The primary work that we do that is relevant here is what we call “health technology assessment” (HTA). Health technology assessment is a set of methods used to evaluate the clinical effectiveness, applicability, and economic impact of a health technology. As economists, we are acutely aware that every dollar spent on one patient is a dollar we can’t spend caring for another, so we’re ultimately interested in knowing how every dollar can achieve the most health benefit, or what we call cost-effectiveness. So, when we do HTA, we are trying to identify what of several competing options is going to be the most cost-effective to the purchaser. HTA doesn’t have to be conducted in a single-purchaser health system environment, but that is typically how we apply it.

Within an HTA, we first need to understand the current standard of care that the health technology in question would affect. Specifically, we want to understand what the current standard of care costs and what the associated health outcomes are. From there, we then need to know where the new health technology fits within that standard: is it supplementing the current process, or is it replacing some part of it? What are the implications for total costs to the health system and to society, and what differences are expected in health outcomes for patients, caregivers, and other parts of the health system?

For example, we might see benefits in terms of reduced resource use within a hospital setting, which might not necessarily be fully captured within a health system costing model, but would still be highly relevant to hospital operations. As another example, if you can improve ER wait times, that doesn’t necessarily change the overall cost of running that ER in a substantial way, but in terms of efficiency, throughput, and staff well-being, it might be very influential. So those are additional outcomes that might be relevant. Again, that all falls into the broader categories of measuring costs, measuring health outcomes, and measuring functional or system-level outcomes that may be important to decision-makers.

From there, we build a simulation model, which is a mathematical representation of the current standard of care, in which we simulate patients moving through that process and through the healthcare system. We capture variability in health outcomes and variability in costs based on patient characteristics, health system characteristics, differences in physician practice patterns, and differences in patient response to treatment or diagnostics. In other words, we account for heterogeneity within the population and within health system performance. We then create a base-case model that represents the current state of affairs—the current version of the world. This allows us to estimate what it currently costs the health system to deliver care and what patient outcomes look like under the status quo.

Starting from that base case, we then introduce the new health technology into this simulated version of the world. To do this, we draw on a variety of data sources: clinical trials, pilot studies, observational data, or in some cases early effectiveness estimates. We run the model and observe how outcomes and costs change relative to the base case.

We run the model many times to account for heterogeneity in outcomes and uncertainty in the treatment effect and other model parameters. This gives us a full picture of the expected health outcomes, system outcomes, and costs. From there, we conduct the economic analysis. At this point we get into cost-effectiveness analysis or cost–benefit analysis, which are structured ways of translating those differences in costs and outcomes into decision-relevant metrics, based on what we understand the priorities of the decision-maker to be. This allows us to quantify the trade-offs involved in adopting this technology versus not adopting it, and to more clearly understand which version of the world we would prefer to implement. Of course, this usually involves uncertainty, so we need to explicitly account for that as part of the decision-making process.

That overall process is the health technology assessment. We present that information to the decision-maker, and the decision-maker then makes a decision. Sometimes the cost-effectiveness results are clear and it’s a relatively straightforward decision, but more often the decision is complicated by evidence uncertainty, heterogeneity, and data limitations, all of which introduce decision risk. Part of the process is helping the decision-maker understand that risk, understand their options, and identify whether there are ways to mitigate that risk—for example, through phased implementation or conditional adoption.

KB: How do you define “optimal” when you evaluate standards of care?

SVK: It depends on what kind of intervention we’re trying to model. Let’s take a really simple example: a diagnostic tool. There’s no shortage of new health technologies that aim to improve diagnostic efficiency and diagnostic accuracy. There’s also a ceiling on how effective a diagnostic tool can be—you can only be so accurate and you can only be so fast. So that’s a good example of what we mean by “optimal”: the upper bound on achievable performance, given technological and clinical constraints.

A more complicated example is something like diabetes, where there are multiple points in the care pathway where you can intervene. Once somebody has diabetes, there are many ways we can treat and manage those patients, and we can model that standard of care. But we could also intervene earlier through prevention, screening, or better self-management to prevent cases from occurring or to delay progression. The optimization question becomes: what combination of interventions across prevention, management, and treatment produces the greatest health benefit for the population we have and with the resources available?

When we model this, “optimal” usually means the most efficient allocation of resources across those different parts of the pathway—where an additional dollar spent produces the most health benefit. Preventing a case entirely often represents something close to an upper bound on potential benefit, because you avoid all downstream costs and health losses associated with that case. But in reality, the optimal solution usually involves some mix of prevention, monitoring, and treatment, and the model helps us understand where along that pathway investments have the greatest value.

We can’t always calculate a perfectly optimal system in a literal sense, but we can usually identify which interventions are likely to produce the greatest benefit at the margin, and that’s usually what decision-makers actually need to know.

KB: Clearly the process you’re describing is only as good as the data that informs it. Understanding that the quality and quantity of available data will reflect the stage of development of a new health technology, where is that data typically coming from?

SVK: Evidence collection is the biggest part of any health technology assessment, and usually we’re combining data from many sources. At a minimum, we are collecting evidence about the health technology itself, which ideally is coming from trials. From that, we might get data on the effectiveness of the technology. We also need data comparing those results to some standard of care or alternative technologies that are already in use.

That can get complicated, because sometimes a trial conducted in another jurisdiction, like the United States, may compare the technology to a standard of care that is different from the standard of care in Canada. They might be using different drugs, different technologies, or different clinical pathways. In these cases, there is some extrapolation and statistical adjustment required to make sure we are comparing apples to apples and that the results are applicable to the jurisdiction where the decision is being made.

Then there’s the system-level data, which we try to collect from administrative data bases or fee schedules. We then have to evaluate that data, clean it, and transform it into something that allows us to simulate the care pathway and the patient population. Often, administrative data has gaps that we need to fill—we may need to impute missing data, forecast certain parameters, or better characterize populations that are not well captured in the data but still need to be represented in the model. Sometimes this involves making assumptions informed by alternative data sources, including qualitative or observational data about those populations.

Finally, we may also use data from the published literature that looks at things like patient experience, physician practice patterns, and health system implementation in different settings. Health systems are not implemented uniformly, and outcomes can differ significantly depending on factors like urban versus rural settings, socioeconomic characteristics, and real-world challenges like differences in clinician or patient adherence to guidelines. We need to account for those factors if we want the model to reflect the real-world population and to support equitable decision-making.

The fact that there are so many different sources of data that need to be brought together is actually why simulation modelling is used in the first place. If all of the data came from a single source, we could often answer the question using a simpler statistical analysis. It is because we have to combine clinical data, administrative data, literature, and assumptions about how the system operates that we build a model of the care pathway and the population, and then run that model many times — what we call probabilistic sensitivity analysis — to generate a range of possible outcomes. That allows us to understand not just what we think will happen, but what could happen under different assumptions and accounting for uncertainty in our data, and only then can we draw conclusions for decision-makers.

KB: I understand you’re building an early healthcare assessment tool to be able to start making predictions about the potential value of these healthcare technologies earlier in the process. How you mitigate the inevitable data quality issues that come with the early technology in general?

SVK: I would describe early health technology assessment (eHTA) not necessarily as a single tool, but rather as a set of methods. We’re taking the same HTA framework but applying it earlier in the product development process, recognizing that at earlier stages we have less data available, which means there will be higher levels of uncertainty in our conclusions. That introduces a unique set of challenges, but when we’re talking about early-stage innovation, there’s a bunch of valuable insights eHTA can still offer.

We still assess the current standard of care and the current care pathway. What we don’t necessarily know at that stage is the true treatment effect of the new health technology. Often, the technology is early enough in development that we don’t even know exactly where it will fit in the health system. For example, it might ultimately be best suited to a specific sub-population, a specific indication, or a specific point in the care pathway.

With that in mind, what we do in eHTA is develop analyses that are not meant to definitively conclude whether the technology will be cost-effective — it’s too early for that. Instead, we focus on answering a few key questions.

First: given the current standard of care, what is the unmet need?

Another way of asking this is to what extent the current standard of care is not optimal compared to best achievable outcomes or ideal patient health. If we are already achieving very good outcomes, then there may be limited room for improvement, and therefore limited value in introducing a new technology. We can quantify this as “headroom” — both therapeutic headroom, how much health improvement is possible, and economic headroom, how much are we willing to pay for that improvement. This helps determine whether the unmet need is large enough to justify the cost and complexity of introducing a new technology.

Second: given the current standard of care, how much more effective would a new technology have to be, at the point it is ready for adoption, for it to be considered cost-effective?

At this stage, we may not yet know how effective the technology actually is — it may not have been trialled yet, or it may still be at the prototype stage. So instead, we ask: given the current standard of care, what level of improvement would be required for any new technology to be considered cost-effective at a given price? This is what we call threshold analysis. We essentially map combinations of cost and effectiveness and identify the boundary at which the technology would be considered cost-effective. That allows developers to see whether the required improvements are realistically achievable and whether the technology is likely to be economically viable before they invest heavily in development and trials.

The third type of analysis applies when a technology is closer to the clinical trial stage. Clinical trials are extremely expensive, and it is not necessarily sufficient to run a trial, publish the results, and present those results to a decision-maker. A purchaser may say: “You’ve shown that the technology is effective, but you haven’t shown whether it is cost-effective, and you haven’t estimated the budget impact.” In other words, decision-makers require evidence that goes beyond clinical effectiveness alone.

So the question becomes: what evidence does the decision-maker actually need in order to make a decision, and what data should be collected during the trial to support that decision? For example, we might know that a technology improves life expectancy, but we don’t yet have good data on quality of life gains from better patient outcomes, which means we cannot calculate cost-effectiveness. That tells us that collecting quality-of-life data during the trial is critical. In other cases, quality-of-life data may already be well established, and instead it may be more important to collect data on adverse events, diagnostic accuracy, or resource use. Wherever the biggest evidence gaps are that lead to evidence uncertainty usually lead to the highest decision risk. We can price that risk.

We call this value of information analysis — estimating the value of collecting additional information, in terms of how much it improves decision-making. It helps determine whether it is worth the cost of collecting certain data as part of a trial.

So those are three broad types of eHTA, and they correspond to different stages of technology development. At the earliest stage, when you may not even have a prototype yet, eHTA is mainly about determining whether there is sufficient unmet need and potential value to justify investing in development at all. At the next stage, when you have a technology concept, threshold analysis helps determine whether it is economically feasible — whether it could be effective enough and inexpensive enough to be adopted. Finally, when you are preparing for trials and evidence generation, value of information analysis helps determine what data you should collect so that, after the trial, you have the evidence needed to support adoption. That helps ensure that you only need to run one trial, not two or three, to generate the evidence required by decision-makers.

KB: I am seeing a lot of interest from policymakers in identifying research that addresses a societal need. Could eHTA be flipped on its head and used to map out where there’s lots of economic or performance headroom? If so, it could be a key input into how health research funding could be more efficiently allocated.

SVK: Exactly right. These methodologies can be, as you say, flipped on their head: instead of starting with a solution and asking what problem it solves, we can start by identifying the biggest problems for which we need better solutions. The same methods can be applied in that direction.

We can use headroom analysis, threshold analysis, or value of information analysis to assess a health domain and identify where the largest sources of unmet need are, where the largest performance gaps exist between current care and best achievable outcomes, and where the biggest areas of uncertainty are that currently limit decision-making in the health system.

In the case of value of information analysis, we can actually quantify how much a health system should be willing to invest to close the gap between the current standard of care and the optimal standard of care. By quantifying that, we can help guide funding agencies, innovation accelerators, and health systems toward the areas where even small improvements in health technology or care delivery would translate into large health or economic gains. In other words, it provides a way to prioritize research and innovation funding based on potential system-level value, rather than just scientific interest or technological feasibility.

We’ve definitely seen excitement from policymakers on that front. The challenge from their perspective is that there are many competing health priorities, so the real question becomes where to start and how to prioritize across disease areas and interventions.

KB: Based on what you’re seeing from results of applying this process so far versus how you’re seeing actual healthcare research spending being allocated, do you have a sense of the kind of efficiency improvements we could see were our research funding being more effectively allocated at the front end?

SVK: It’s a complicated question. I think the simple answer is that, in my experience, when I look at how health system priorities are set, it’s largely done by looking at the biggest ticket items. This usually means that conditions with the highest prevalence and the highest healthcare system cost get the most attention. Which may seem logical. But once you look more closely, you realize that there are significant gains to be had in these less common health conditions that are currently given much less attention. There is enormous headroom in these areas.

Note that these are not necessarily rare conditions, because rare conditions are often already very expensive to treat and therefore do receive attention. Rather, I’m talking about things like preventive interventions and population health measures aimed at improving the overall health of the general population. Those areas are often under-invested in, even though they can have substantial downstream effects: they prevent late-stage disease and delay disease progression, which can produce very large long-term benefits. From an economic standpoint, that is often where some of the largest gains can be realized, but in practice I see fiscal pressure and immediate budget impact driving a lot of prioritization.

I’ve seen that there’s a lot of patient advocacy that informs priorities. This is a good thing — it’s important to a democratic health system. But it’s often the case that when a new technology is developed, suddenly there’s a tremendous amount of advocacy about that solution and about the importance of implementing that solution, and so all of a sudden a disease gains a level of attention and a level of prioritization because we because we found a new solution to it, which puts priority onto a condition based on the existence of a new technology rather than prioritization technology development based on where the most need for health system investment actually lies.

KB: Coming back then to the innovator and the researcher side of the equation, what should innovators and researchers be thinking about before having a conversation with you and with IHE?

SVK: The most common blind spots we see when working with innovators are, first, that the innovator doesn’t know who their purchaser is, and second, that they don’t know what their purchaser values. It’s not really the innovator’s fault that they don’t know these things. Our health system isn’t particularly transparent about how decisions are made. The process differs by province, sometimes by health authority, and the regulatory and decision-making processes are evolving, with new approaches being introduced over time. We don’t make it easy, so it’s understandable that innovators often don’t know these things.

The main thing I would say is that every innovation is different, every pathway into the health system is different, and the decision-making environment is often difficult to navigate from the outside. That’s really where we see our role — helping innovators and researchers understand that landscape, understand what evidence is going to matter, and think through how their technology can realistically fit into the system. We’re always happy to have those conversations early, because the earlier you start thinking about these questions, the better your chances are of generating the right evidence and ultimately getting your innovation adopted.

Key Takeaways

Recent changes to provincial research funding mandates make clear that Canada is starting to reorient its investment in research toward projects and technologies that respond to a clear societal need. This is a non-trivial task, and funding agencies have mostly used the existence of a partner organization willing to co-fund the research as the primary signal of societal need. This is reflected mainly in grant eligibility criteria, for example NSERC Alliance and I2I. However, this approach leads to a very narrow definition of societal need, as there are many challenges for which no champion organization exists yet. Focusing entirely on private sector involvement leaves opportunities for valuable and impactful research on the table. Sasha’s commentary on eHTA makes clear that we have a much larger toolkit at our disposal for identifying research worth funding than basing it entirely on existing private sector involvement. Where health technology is concerned, it is actually possible to optimize our investment of research resources in a way that maximizes impact. As Sasha puts it:

“[I]stead of starting with a solution and asking what problem it solves, we can start by identifying the biggest problems for which we need better solutions. […] In other words, [eHTA] provides a way to prioritize research and innovation funding based on potential system-level value, rather than just scientific interest or technological feasibility.”

Part of why this works in the healthcare space specifically is that the process of adoption of novel technologies is (at least somewhat) systematized, through the regulatory pathways that govern whether and how new technologies become part of the public health system in Canada.

Health regulation is often viewed by early-stage innovators and investors primarily as a hurdle to be overcome. It is expensive, complex, and disproportionately impacts new startups that do not yet have the internal processes experience to navigate it effectively. Sasha’s commentary makes clear that it can also be the basis for effective decision-making and can inform optimization of both public and private sector health tech investment. By creating predictable pathways to impact, regulation contributes to making it possible to assess the potential value of technologies much earlier, in ways that simply would not be possible in a less regulated space.

Starting from the idea stage, eHTA can be used to progressively refine the direction of research and decide at each step whether and how to move forward:

“At the earliest stage, when you may not even have a prototype yet, eHTA is mainly about determining whether there is sufficient unmet need and potential value to justify investing in development at all. At the next stage, when you have a technology concept, threshold analysis helps determine whether it is economically feasible — whether it could be effective enough and inexpensive enough to be adopted. Finally, when you are preparing for trials and evidence generation, value of information analysis helps determine what data you should collect so that, after the trial, you have the evidence needed to support adoption. That helps ensure that you only need to run one trial, not two or three, to generate the evidence required by decision-makers.”

This is not to say that regulation is inherently a good thing — simply that, if it is consistent and well-understood, it can be inform high-quality decision-making.

It is apparent, though, that Canada’s regulatory landscape in the health space suffers from the same fragmentation that plagues every other aspect of its innovation pipeline:

“Every province has its own process, and the level of development and maturity of that process is highly variable. In some cases, the decision is not even made at the provincial level; instead, it may be made at the health system or regional health authority level, with organizations making adoption decisions independently. Even when it is a provincial decision, the kinds of evidence being presented and what the pathway toward adoption looks like can vary substantially, and sometimes the process is not especially well defined, even within a given province or territory. It might not even be clear to policymakers themselves what the exact process is in a given case.”

Obviously, this is not a landscape that can be reasonably understood by innovators and researchers that have not navigated it before. However, it is one that can and should be used by research funding agencies to identify health research that responds to a societal need, and one that can be the basis of a tight feedback loop between research funders, the researchers applying for funding, and the innovators taking that research beyond the lab:

“If we are already achieving very good outcomes, then there may be limited room for improvement, and therefore limited value in introducing a new technology. We can quantify this as “headroom” — both therapeutic headroom, how much health improvement is possible, and economic headroom, how much are we willing to pay for that improvement.

We essentially map combinations of cost and effectiveness and identify the boundary at which the technology would be considered cost-effective. That allows developers to see whether the required improvements are realistically achievable and whether the technology is likely to be economically viable before they invest heavily in development and trials.

[…]

Wherever the biggest evidence gaps are that lead to evidence uncertainty usually lead to the highest decision risk. We can price that risk.”

This kind of analysis is critical both for funders and innovators. I cannot count the number of times I’ve engaged with an early-stage health tech company excited about a new technology that have already started to build a company without having given any thought to the economics of what they are trying to build.

On the funder side, the alternative is a more reactive approach in which research breakthroughs lead research funding investment decisions:

“[I]t’s often the case that when a new technology is developed, suddenly there’s a tremendous amount of advocacy about that solution and about the importance of implementing that solution, and so all of a sudden a disease gains a level of attention and a level of prioritization because we because we found a new solution to it, which puts priority onto a condition based on the existence of a new technology rather than prioritization technology development based on where the most need for health system investment actually lies.”

While not a problem in and of itself, it is unlikely that a reliance on serendipity will lead to economically optimal use of scarce resources. Effective research funding policy must consider both focused discovery and opportunistic engagement with unplanned breakthroughs.

Not only does eHTA help decide what is worth pursuing at the research funding stage, it makes what is pursued more likely to succeed at each step of the process while reducing cost and risk en route. It’s not about picking winners (many early health technologies can and will still fail to achieve their potential impact for all kind of reasons unrelated to their economic and technological merits), it is about being thoughtful in matching spending to real need so that when there are successes, they are impactful at a societal level.

As we shift our research priorities toward addressing unmet Canadian challenges, analysis like eHTA should inform the investment of at least a portion of our research funding. The existence of a private sector company is not the only signal of demand for innovation — and with a reasonable expectation that technologies will be cost-effective ahead of time and a clear understanding of the risks, research funders allocating resources informed by eHTA can be confident that the private sector will step into the niche at the appropriate time in the development process.

Many thanks to Sasha for sharing his insights.

A recent study of what the authors call “university venture capital” (UVC) in North American and Europe provides an interesting point of convergence for several threads that have been floating around the innovation ecosystem over the past several weeks. Before reading the rest of this article, I suggest at least skimming “University venture capital in Europe and North America: Evidence, models, and EU policy implications”, an article by Thomas Crispeels, Emiel De Buyser, James P. Gavigan, and Ramón Compañó. This work contains a wealth of data and policy insights from across a large subset of OECD countries, and is well worth a read on its own.

They frame the problem clearly:

“[University spin-offs (USO)s] often encounter significant barriers in accessing early-stage investments by Business Angels and Venture Capitalists (VC). Their dependence on intangible assets, early-stage technologies, and the absence of a commercial track record make it difficult for investors to evaluate their viability. This information asymmetry leads to a persistent funding gap, as traditional VC firms tend to favour more mature ventures with clearer market potential.”

The results strongly support a recent report by the National Angel Capital Association (NACO) that demonstrated that Canada is missing 141M and $181M USD in pre-seed and seed-stage investment capital, respectively, relative to the United States, resulting in $66B in foregone economic value creation. Contributing to that gap, Crispeels et al. show, is that Canada sits in last place for total UVC among countries surveyed (see Figure 7, reproduced below).

This total number hides a heterogeneous domestic UVC landscape. Several Canadian universities have investment arms that are of a size consistent with the global medians in terms of total assets under management, as discussed in detail in my article covering such funds. Acknowledging the existence and importance of those outliers, the fact remains that Canada lags the world in this area, a problem I believe is foundational to our lagging productivity in general. A careful read of this work, informed by the recently published simulation model on effective investment strategies for emerging technology, offers clues as to structural reasons for the gap.

In this article, I connect the dots between this paper, my recent simulations of venture philanthropy, my survey of Canadian university deep tech funds, NACO’s findings of missing Canadian early-stage capital, and several university-related topics about which I have written in the past. Combined with NACO’s report, this work is confirmation of several “weak signals” that I have been writing about for some time, and a call to action for Canada both to address these gaps and to commit to collecting the data needed to be able to proactively identify them in the future.

Weak signals

Innovation arises as the result of accurately identifying and acting early on weak signals of impending change. One such signal exists in this paper, as well as in my own previous research on the topic: Crispeels et al. observe a trend toward larger and larger structures in a small but growing minority of UVC funds:

“[W]e observe that the amount of UVC available in Europe is growing, with some notable trends: the emergence of multi-university UVCs attracting larger investments, the occurrence of international multi-university UVCs, and the evolution towards a broader investment scope (e.g., investing in companies that collaborate with universities or in-license university intellectual property rights).”

Figures 9 and 10 in particular, reproduced below, show a small but growing minority of multi-university investment funds. It appears that several universities, across Europe in particular, are pooling their investment resources.

As I was conducting the interviews for my previous article, it came quickly clear that in cases where universities had iterated at least once in terms of their approach, the second attempt usually involved an entity that was at a greater structural distance from the university, and that had an expanded investment scope, as compared to the first.

By way of example, the fund with the longest track record that I surveyed (Velocity Fund in Waterloo) has actually gone through several iterations, ending up with an almost entirely separate entity from the university (the university is an LP in the latest fund) that invests globally. UCeed in Calgary, (built and led by John Wilson, who previously had a hand in building the UK approach), is managed by a wholly-owned but arm’s-length subsidiary of the university, and has a provincial scope for some of its sub-funds. While not a multi-university approach in most cases, the trend toward larger and larger investment scope is the same.

Scale matters

In light of recent results from simulations of investment in university-affiliated companies, this trend should not be all that surprising. The key takeaway from that work was that the single most important determinant of long-term success is scale. Where emerging technology is concerned, it is more important to make sure that you do not miss the homeruns than it is to ensure success with every investment - median returns are dictated entirely by the presence of outliers in your portfolio. Combined with the fact that it is all but impossible to predict which attempts to commercialize research will be successful at the founding stage, and it becomes clear that UVC, like any emerging technology investment play, is primarily a numbers game. Seen through this lens, the trend toward larger and larger investment scope that we are seeing in Canada, and the trend toward multi-institution and even cross-border UVC funds in Europe, makes perfect sense.

Canadian universities are at a crossroads, with recent shifts in policy applying pressure toward delivering the “third mission” of universities to demonstrate direct societal impact (and specifically, economic impact) of their research. All of these pieces of evidence point in the same direction: a broadly scoped fund with pooled resources can make more investments, while most single universities alone will struggle to achieve the necessary scale through no fault of their own.

There is a subtle nuance here that is worth spelling out. While the combined effect of many disconnected UVC funds would be the same at the ecosystem level as a pooled fund of the same total size, individually those 10 disconnected funds will experience extreme variance and will individually perform very differently, with some succeeding and some failing, exposing individual institutions to a much higher degree of risk and jeopardizing the long-term impact of the while enterprise some universities drop off, leaving their spinouts without support and driving a heterogeneous landscape of research commercialization effectiveness. A larger, combined fund approach serves both to smooth out this variance, both geographically and at the level of fund performance, to increase the median investment performance, and thereby reduces individual university risk exposure.

The lessons for Canada

This tendency toward consolidated UVC is rational. It is reflected in a recommendation also made by the European Commission. As noted by Crispeels et al.:

“The European Commission […] contains a number of measures to boost the creation of innovative firms in the EU, targeting universities. One measure refers to support cross-border networking and collaboration between leading hubs rooted in strong university ecosystems.”

Of course, this is not the whole story. If emerging technology and research commercialization is a numbers game, then dealflow is critical, and the lack of focus on UVC in Canada also reflects a “chicken and egg” problem, wherein researcher-led entrepreneurship is not necessarily built into the culture of all Canadian universities (though there are some obvious exceptions), resulting in lower dealflow that in turns reinforces the difficulty involved in building effective single-institutions funds. This vicious cycle is examined through different lenses in my article on the importance of embracing risk in early-stage innovation, as well as on embedding entrepreneurial culture in university faculty. The European Commission also addresses this issue:

“A second [measure recommended by the European Commission] is to develop a blueprint for an academic career development framework that rewards research commercialisation [sic] activities, including considering staff mobility between the university and industry in academic staff evaluation and promotion criteria. Rewarding faculty members involved in the commercialization of research results and spin-off creation increases their propensity to contribute more to spin-off formation and knowledge transfer activity.”

The PTIE movement in the United States, which is readily adaptable to Canada, gives us a concrete roadmap for addressing this point.

While great progress has been made in several Canadian regional and provincial ecosystems with respect to university spinout investment, there remains an enormous amount of Canadian research that lacks reliable pre-seed funding for commercialization, as evidenced both by this paper and NACO’s findings.

Canada has a lot to learn from our European allies, and the findings of this paper will be an important consideration as Canadian universities (and Canada generally) responds to the early-stage capital gap and to shifting societal expectations of its universities. The important takeaway from the point of view of the broader Canadian innovation ecosystem is that weak signals are critical to getting ahead of global trends, and detecting them is impossible if we are not collecting and critically evaluating data on our innovation performance, at every level.

As we enact sweeping changes to our approach to innovation, now more than ever Canada needs to make collection and coordinated analysis of effective, long-term performance metrics a matter of course across all levels of our innovation ecosystem. This will enable us to detect these weak signals early so that we can proactively address gaps, rather than react to their revelation after the damage is done and the signal is no longer weak.

Simulating venture philanthropy

This post presents a project that has been in the works for a long time, and one that combines everything I’ve written about risk tolerance, venture philanthropy, and the Simple Agreement for Innovation Licensing (SAIL) into a single, cohesive framework for investing in emerging technology. Using data from PitchBook, I built a simulation model that enables me to back-test the performance of venture philanthropy as a means to invest in emerging technology. A pre-print of that article is available at the link below.

Read the pre-print

The results in the paper validate the thesis that investing in emerging technologies is primarily a numbers game, and that attempting to pick winners actively drives underperformance. It shows that there is good reason to believe that venture philanthropy will be self-sustaining if (and only if) it operates at the right scale, and it explains clearly why neither the public nor private sector can address the problem alone—risk must be shared for this to work. It also shows that investing in emerging technology is worth doing even if it fails more than 96% of the time.

Putting risk tolerance under the microscope

It’s now well-established that Canada is great at research but struggles to realize socioeconomic value from the results. For as long as CanInnovate has existed I have argued that this problem is structural, and is rooted in risk-intolerance. The thesis is not all that complex, and relies on three assumptions:

1. that emerging technologies follow an extremely skewed distribution of value in which a minority of technologies produce a majority of the return;

2. that it is practically impossible to predict which technologies will ultimately be in the valuable minority in the early stages; and

3. that the value of the small number of successes more than offsets the cost of the majority of failures.

The first two assumptions are, I think, well-established as truth at this point. The third one requires validation. The purposes of this paper were to demonstrate, using real-world performance data from university-based startups around the world, that the third assumption holds true, and to identify precisely why neither the public nor private sector alone have yet addressed the failure to bring emerging technologies to market.

If these assumptions are sound, it follows that the only “winning” move is to invest in almost everything, accepting the risk that most will fail. A number of previous simulation models have demonstrated convincingly that in these conditions, all that matters for investment performance is that you do not miss the valuable minority. You can find a simple model that demonstrates this here, and a more sophisticated one here. Their conclusions are broadly the same: performance depends heavily on the number of investments made, requiring a large number of investments and patience to cut through the noise of high failure rates and long development cycles.

Risk intolerance is a self-fulfilling prophecy of underperformance: risk-intolerant investors try to pick winners and as a result make only a few bets, increasing the chances of missing a homerun. Given the rarity of truly valuable technologies and the second assumption above, this undersampling effectively guarantees that they are missed, which causes their investment portfolios to underperform. This then becomes the rationale for risk-intolerance in the future, trapping an ecosystem into a negative feedback loop that results in a shallow pool of investment opportunities. Recent research featured at capture the problem simply:

Pool quality acts as a ceiling or constraint - you can’t pick winners that don’t exist in your pool.

Thus, a small improvement in pool quality yields greater returns than a much larger improvement in picking skill, as picking skill compounds on pool quality to produce alpha.

Put another way, your ability to pick well only matters if there are good options to pick from.

The only way out of this cycle is to embrace risk and accept that a majority failure rate is a positive indicator that we are investing broadly enough to make sure we are not missing the valuable opportunities. Risk and failure are key elements of strategy, rather than things to be minimized.

The need for Canadian data

Aside from the conclusions and caveats noted in the paper, there is one task that remains: creating a version of the simulation model that is specific to Canada. The model built for this paper is based on startup companies from the US, Europe, the UK, and the Nordic countries, with a majority of the companies being based outside North America. While the innovation ecosystems in the EU and UK are less different from Canada than, say, that of the United States, this still presents an obvious issue when trying to make quantitative predictions about the Canadian ecosystem.

To address this, more Canadian data is needed. Specifically, data on the founding date, the complete fundraising history, dilution taken in each round, and the eventual outcome (if it has been resolved), of Canadian startups that set out to commercialize publicly funded research. If you know of a source for any of this data (or even just a subset of it), please get in touch.

Many thanks to Innovation Support Services at the University of Ottawa and to Intellectual Property Ontario for funding this research.

This week I interviewed Stefan Leslie, the CEO of Research Nova Scotia (RNS). RNS is an arm’s length, board-governed organization that deploys research grant funding on behalf of the provincial government. Notably, RNS was recently given a new mandate to fund research that will drive provincial economic growth, and they have just finished their first call for proposals under the new strategic plan they developed in response.

My goal in this interview was to learn how Stefan Leslie and the RNS team approached rebuilding their strategic plan to address their new mandate, in the hopes of coming away with a clear articulation of what it means for research to serve the needs of society that can be generalized to other Canadian funding agencies. Stefan’s commentary is insightful, and reflects a deep understanding both of the research he supports and of what it takes to move the results of that research beyond the lab.

While Nova Scotia may be leading the charge, I strongly suspect that similar changes are coming to the rest of the provinces in the near future. The strategic plan developed by Stefan and the RNS team will no doubt serve as a useful blueprint for both provincial and federal policy makers that are tasked with connecting research to societal impact.

Your email client will probably truncate this post. My key takeaways are presented at the end, so be sure to read the web version if you want to get the whole story. Many thanks to Stefan for taking the time to share his process.

Interviewer’s note: Stefan Leslie approved the final version of the section entitled “Interview with Stefan Leslie” and had editorial input on that section, with the option to rephrase and expand on the ideas discussed in the interview without changing or removing any intended meaning. The key takeaways presented at the end are my own commentary, and do not necessarily represent the views of Stefan Leslie or Research Nova Scotia.

Interview with Stefan Leslie

KB: Tell me about yourself. How did you find your way to leading Research Nova Scotia?

SL: I actually came into this from managing fisheries: 12 years in Canada and 4 in New Zealand. When I say managing fisheries, I mean for the government authority, in Canada for DFO and in New Zealand for the Ministry of Fisheries. This included stock assessments and meetings with fishermen in the local fire hall and all that kind of stuff.

What was interesting about coming from that direction was that it is a very science-based activity where you have a very direct connection between the research and science that’s being done and resulting management and policy decisions. That includes fundamental research work around how stock dynamics operate, and then very applied stuff like figuring out how we can exploit this stock while remaining sustainable. I also recognize, having worked in that environment, that you cannot optimize just based on the outcome of a scientific process. People are involved, and they have to come up with the trade-offs and develop the normative choices that come with competing interests around science and research. I also came to recognize my own limits as a public servant and both the limitations and the advantages of working in that kind of environment.

That background then led me to lead a Network of Centres of Excellence, which is a program that no longer exists in Canada as it was cancelled four or five years ago. This was a federally funded program embedded in a host organization, typically a university, that did very purposeful science. It was to pursue certain outcomes that were of importance to the country, develop high-quality work, and also train the workforce, engage in knowledge transfer, and all the rest of it. But it got me into the academic environment and illuminated the value that academic research, or research that happens in the universities, can bring across a whole bunch of different dimensions of Canadian society and the economy. It also highlighted for me the value of an arm’s length, non-profit-style board-governed world.

When Research Nova Scotia was started, and the conversation in the province of Nova Scotia really kicked off around 2015 or 2016 and had about a 3- or 4-year gestation period, the idea was to create an arm’s length organization devoted to funding “purposeful research” with research as the basis for this pathway to prosperity and societal improvement. That was the sentiment in that time, in the pre-pandemic world. So as they wished to launch this organization, I was very interested because I thought that it was an unexplored dimension in the provincial interest in science.

I wanted to explore what this meant that’s different from straight economic or federal interest, because the dominant theme of research is often driven by federal considerations, but the provinces are actually interested in different things. They of course are managing universities, but also research hospitals or healthcare systems where health research happens. So there’s a provincial jurisdiction element. Provinces are responsible for the delivery of the majority of public or social services that you and I and everyone else need and enjoy and comment on on a daily basis: the education system, the health system, transportation, economic development, and so on. As a result, provinces have a real interest in the outputs of research, and are asking the question as to what are we actually learning from research that we can then apply in a direct way.

KB: It sounds like part of the goal of Research Nova Scotia, even before the recent mandate change, was to fund research that would have a direct socioeconomic impact. Is that accurate?

SL: From the outset, I should probably say that in Nova Scotia, and probably other provinces, there have been various approaches to supporting research since at least the 1930s. Research Nova Scotia, although it was created in 2019, is at the end of a long train of a variety of different options and selections and ideas around how the province can support research. This is attempt 7 or 8 at this particular problem, so we’ve tried every sort of model. The immediate predecessor organization had really focused on health research and leveraging federal funds primarily through the Canada Foundation for Innovation. So there really always was this purposeful element of connecting the public investment to issues that matter here. What needed to matter was the research outputs: what we learned, beyond just what we were funding, research excellence, and the people trained. What value could we actually extract from it?

For the first five years more or less of our existence, the province had established research priorities in regulation itself. They were fairly broad and enabling: inclusive economic growth, a healthy population, and a strong healthcare system. Quite broad, but nevertheless indicative of a direction that they wanted to take. Our first approach within that purposeful or intentional model was to develop an outcomes- or missions-based outlook in which we would define fairly broad but meaningful outcomes and we then invited researchers to come and creatively articulate how they could deliver those outcomes. We were quite open for researchers to provide insight into how they could devote their skill and craft in service of achieving that on behalf of the province.

There are diverse ways you can divide up the types of research that exist. You can break it into basic or fundamental and applied and experimental design, like the way the OECD does. Sometimes you can throw in curiosity-driven or investigator-led research. I don’t think necessarily that these are all that helpful as distinctions. We were never in the game of funding discovery for the sake of pursuit of knowledge, as that was not the intent at the provincial level. There was a desire to use scientific tools to fill a particular need, however, and that can call upon the need to engage in fundamental research. If that’s the limiting factor to achieving whatever societal outcome you want then that’s what you ought to be pursuing. We looked at it from the perspective of identifying the “pull” and then deriving the research need from that, regardless of whether it is applied or fundamental by that set of definitions.

As you pointed out, though, there has been a significant shift in what Research Nova Scotia has been asked to do, which has been underway now for the past about eight months. It was initiated by a change in the legislation that underpins what Research Nova Scotia is here to do, in the spring sitting of the 2025 legislature. Many things are still the same: we’re still an arm’s length organization, we’re still independent, we still have a board of directors, we’re still researching areas of provincial importance, and we are not, consistent with iterations of the research endeavour in this province, performing research itself. Instead we’re coordinating and funding others. We are not intervening on choices around methods, nor are we controlling outputs, none of that stuff that flared up during the Harper years that has sparked discussion around political control over science, none of that is part of it. But there are a number of really important changes that have happened.

The first is that under this revised legislation, the Minister of Advanced Education, which is the part of the crown to which I have a reporting relationship, now has the discretion to establish research priorities. So it’s been taken out of regulation and given to the minister. You could say that that’s just taking your coins from one pocket and putting it into another, because it is still a government process, but the regulatory process and a minister’s process is really quite different, or at least it can be, since the minister now has in their sole discretion the ability from time to time to determine those research priorities. That’s a critical difference for us.

The second is that the minister did exercise that discretion very early on under the revised mandate by giving us an overtly economic focus. So we now have to demonstrate advancements across 7 key outcomes which are grouped into two economic areas: economic growth and economic productivity. More specifically, provincial economic growth, and provincial economic productivity. So everything we do, and our authority to spend public money, depends on our ability to connect the research we fund to the achievement of outcomes within those two areas.

Third, which is related but quite important, is that there is an expectation of our ability to connect research outputs to those outcomes in a very direct way. The expectation from the minister is to show how research is contributing to measurable changes in those outcomes. A lot of weight that is held by the word “is” and “measurable”. “Is” really does imply that there has to be that obvious connection and “measurable” means you have to be able to point to some evidence.

The fourth major change is that we have had an adjustment of the priority sectors.I mentioned before that it was inclusive economic growth and healthy population and a healthcare system, and now the effort that we must devote ourselves to has been diverted to three different sectors:

1. Natural resources, climate change and clean energy;

2. Life sciences and health sciences; and

3. Construction and transportation.

So that has been the shift that happened on the 1st of May of last year of 2025.

KB: You’re one of the first if not the first provincial research funding agencies that has been asked to reorient towards economic impact, but I don’t think you’re going to be the last. Tell us about the process you undertook to develop your strategic plan in response, as a guide to those who may be asked to do the same in other provinces.

SL: There was no transition period. It wasn’t as if we were advised that we needed to change our priorities and get there in 12 months. We had the new set of priorities, and our research activities needed to respect them, effective immediately. We devoted the summer to developing a new strategic plan, and in fact the act specifically directed us to create and implement a strategic plan to deliver on our new mandate. That’s good practice anyway, but nevertheless it was helpful to have that pointed to us.

We went through a relatively formal iterative process where we engaged with insiders, got some drafting done, sent material out, and gradually over the period of about three months refined the ideas into a strategy that then went to the board for further refinement and then was approved in October. So approximately six months from creation to adoption phase.

It has three different areas of focus. The first, which almost sounds comical to state, but I think is actually really important, is that we still fund research. The reason I dwell on this is that it matters what people call “research” because it’s actually a term that carries a lot of weight in a lot of different ways depending on the circumstance. If you go back to the Frascati definition of the OECD, it’s got these aspects of novelty and creativity and uncertainty and all the rest of it, and so for us it has to be more than a phrase, which I thought was a really useful distinction point that I learned from this engagement process.

A colleague organization in the province described what they did within a particular sector as “managing scientific activities to resolve knowledge gaps”, and that is a perfectly reasonable definition of what research is, but for us it had to be a little broader. We needed to see absolutely the connection of the research outputs to realizing those outcomes, but we take broader view of research. For RNS, the discovery or the knowledge that’s gained through that research process has to be applicable beyond the immediate circumstances by people who weren’t involved in the initial investigation. This differentiates ourselves from simply supporting R&D, that may happen either in a company or an industrial participant or a government department that has to deliver on a policy mandate.

The second, consistent with that interpretation of what research is for RNS: we adopted a sector-level view. So that our support for solving problems, removing barriers, pursuing opportunities, whatever it is, would be applicable or relevant to a number of participants.

Senator Deacon, back in October or November, put out a report talking about all of the different federal level programs that support R&D, everything from tax credits to direct support, and there’s hundreds of them. Our view was that there’s no point in us simply being number 135 where there had been 134. Rather, we asked how we could incentivize collaboration within industry and make sure there was a connection between industry needs, and other indications that the research will produce economic value, and the research performers, whether they’re at a university or a not-for-profit or even within industry. So we were looking for opportunities to incentivize this collaboration where there are common challenges that have an economic benefit but are broader than just what one company needs in order to sell a particular product, improve a particular process, reduce their cost structure, or whatever it might be.

That leads to the third main takeaway from our strategy. Consistent with the first two parts, it behooved us to create a couple of different pathways to research support. First, we have a convened model which we call Focused Research Investments, which are large, they’re actively coordinated, there’s multiple component or constituent projects that fall out of it, in which we are looking for active engagement with other funders to really advance in an area over a period of multiple years. So you pick a sector for which there is an opportunity to manage a risk, or pursue an opportunity, and you say that, by virtue of systematic and consistent application of money, time, and talent, that you can really advance things along. Second, this is matched by a more responsive model, which enables those who have an idea, or their own challenge that they can cogently describe, to get funding over a shorter period of time toward resolving that challenge. You don’t have to have all that other bells and whistles around coordination here. You do need to be able to articulate how it fits our definition of research, but you don’t have to coordinate a larger scale effort. This is the opportunity for which our first intake round just ended. With that first round, applicants are being invited to propose what they think is important for us to pursue over the long run.

In both cases, whether you’re coming through that convened larger model or the responsive model, you need to be able to demonstrate a plausible pathway to impact. We can discuss the economic model later, but the summary point here is that applicants do have to demonstrate that the need of whoever’s going to apply the research, which has to be outside the research community, shaped the direction of the research that’s occurring. Applicants need to understand that particular pathway. This is different from promising success. Rather, applicants need to demonstrate that they have thought about who and where and when information moves through hands or through different parts of the value chain, to go from the research sphere to the application sphere. Just as important is the demonstration of an effective mechanism or governance structure within the research component that can make adjustments or accommodations as you learn to ensure that your focus is always on that outcome.

We have lots to learn on whether this is going to work, but by taking these two different models and keeping our eye on that application phase, we’re hoping to back that into a model that will allow us to identify and then track success of projects over time.

KB: Your convened model, and in particular the idea of removing the separation between research and industry, has some parallels to a more agile version of DARPA. Was that a source of inspiration?

SL: Certainly DARPA, or Skunkworks is another one that’s commonly brought up, are really great for engineering challenges where it’s easy to identify a measure of success or failure. To use the Skunkworks example, if you are designing a spy plane, you can tightly specify the design and success criteria: it must be able to fly at 90,000 feet and get a resolution of X, for example, and then you have either a spy plane, or you don’t, that meets those design criteria and you can design a research agenda around that. There’s certainly an element of that.

I think what’s different here is both the need to find a way to integrate the existing organizational structure of how research actually happens in this country. Nova Scotia has 10 universities, a community college with multiple campuses, and two or three health authorities delivering healthcare, but we have to recognize that there’s a strong social component to whether something actually happens. It’s not just an engineering challenge.

I’d say the inspiration was partly that DARPA model and the various other ARPA-[add-a-letters]. There are certainly other innovation models, and the US, for all of its challenges around science these days, are extremely creative in trying different models and different ways. Philanthropically supported research is considerably bigger there than it is here (although I think it’s growing here), convergence research, developing the Focused Research Organization model, speculative technology, working on deep tech challenges. There’s the UK version of some of this the Advanced Research and Invention Agency or ARIA that’s been around for maybe two or three years, Deep Science Ventures, various other organizations, etc. I’d say that what is unifying all of these is that they’re experimental and adaptive. It’s not just that they have this connection of a societal or economic or engineering need into the research community. I think what is kind of more special about them is that they are recognizing that you need to adjust the model according to what you’re trying to achieve, be honest about or or look carefully at what it’s taken to succeed, or not, and adjust going forward without trying to determine what the model is that you’re going to hold true to forevermore.

KB: Given the number of unknowns between research and economic impact that may have nothing to do with the quality of the idea or the researchers involved, how are you planning to connect the research you fund to the downstream impact?

SL: This has been an area which has obsessed us since the economic mandate for RNS began last May. But more broadly, this has proven to be a vexatious problem for at least 70 or 80 years. It has been the subject of considerable attention for probably 50 of those years. It is not one that is easily cracked. A few general comments before we get to the more specifics, because I’ve got a couple of different ways to answer that question in a more concrete way.

I think it is well established, and probably beyond contention, that there is a public or a social benefit to public investments in research and development. There’s been a variety of different papers on this, and probably the most recent or maybe best known was by the National Bureau of Economic Research in the US, where Lawrence Summers and Benjamin Jones looked at the social value of research. It is very difficult to find a study that doesn’t give you a multiplier of somewhere on the order of 4-10. For every dollar spent on research, you get four back in social value. The issue is that it really matters where that accrues.

This is the challenge around R&D: you have a high degree of uncertainty in where those benefits are felt, but you have high certainty over where those costs originate. So you have what you could probably characterize as benefits that are diffuse; distant; hard to attribute to an individual supporter, and that are often disconnected from the original investor or research performer, but the costs have all the opposite characteristics. They are coming from a very specific pot of money. Even when R&D is done to improve a very specific thing, say a manufacturing process, it will be on the order of years before you see the application of that knowledge to the improvement of that particular process.

There was a study done not too long ago that looked at the timeframe between publication in a scientific journal, where the research has been “concluded”, and the application for a patent. Not the use of that patent, mind you, just applying it. That’s on the order of three to six years So, if you’re looking for evidence of economic benefit from the activity of research, you have to think about the timeframe of years. The more fundamental the research question, the longer that timeframe is. For basic research, we’re talking probably 20 years.

[[interviewer’s note: Stefan’s numbers line up with the NBER study linked above, in which they suggest that “studies of R&D, product introductions, and product sales suggest quite rapid linkages between up-front costs and peak market payoffs. A total delay of 3.6 years appears reasonable, and a 10-year delay appears very conservative.” (p.14)” and that “”citation network analysis suggests that even more remote basic research investments begin paying off within 20 years.”]].

You can always point to, say, the connection between someone studying the spit of a Gila monster and the production of Ozempic, or you can point to research into how bird beaks were formed through an evolutionary process connecting to methods to improve the aerodynamics of high-speed trains. That all happens, but it is a very tortuous path, and you can make those connections only in retrospect. It’s very difficult to do it looking forwards.

Now I want to answer your question in a more specific way, with two different perspectives. The first is rooted in evaluation. Looking back, how do you demonstrate that research you’ve funded has resulted in value. As I’ve just spent a lot of time describing, you need something that goes well beyond the timeframe of the research project itself. You need to be able to reach well beyond the end of the research, and probably coordinate with others to be able to track it back, and you need to focus as much on efforts and intent as on results. You can’t hold a researcher to account for a failure to realize economic return because there are so many aspects that are going to either frustrate or amplify creation of that economic value over which they have no control, but you can certainly expect that there is effort and intention behind what they do. I think that the metrics you need to look at need to allow you to develop a sensitive understanding about why something worked, or why it didn’t work, so that you can apply those learnings to future investigations.

This is where I need to get to the second area that represents a very different way of thinking about the problem. How do you actually select research projects that have the prospect of economic benefit, given the timeframe and the lack of attribution. Basically, how do we decide what represents the best bet? If you line up 10 ideas in front of venture capital, they’re going to have about a 10 percent success rate. They know what the product is, who the principals are, what the market is, and what the financing looks like, and the success rate is still that low. So, how can you provide information to someone like us, that is being asked to select research, which is well before you get to the point where VC would ever take a sniff, that signals a chance at success?

I think the only way you do that is you look at this in a different way. You have to completely change what you’re looking for. You need not focus on evidence or expectation that whatever comes out of the lab is going to be applied, but instead look for indicators of how the problem within society that needs to be solved is influencing what goes on in the lab. We need to move away from the idea of moving knowledge out of the lab, this “knowledge translation” language we hear so much about, and instead flip it around and consider knowledge translation in the opposite direction: how information from the social side informs the research sphere.

So we need to completely change what we’re looking for, and focus on evidence of how the needs of society or industry shaped the research endeavor. It’s more than “I had a question, I know how something happened and I set about to understand it and hopefully it’ll be applicable”. Instead, it needs to be “are the questions that I’m asking directly inspired by or

The patron saint is like Louis Pasteur. Abraham Flexner, the guy who who started the Institute for Advanced Study at Princeton, doing the purest exploratory research, had all the time in the world for Pasteur, even though he even wrote an article on “The Usefulness of Useless Knowledge“. He talked about Pasteur taking inspiration from the real needs of people and converting those unknowns into fundamental challenges.

We need to take needs and convert those into real problems. Transform the applied into the fundamental rather than take the fundamental and try to apply it. So we need to out that process. I think that the many researchers who do applied research do this in an intuitive sense. They are not cloistered and kind of just come up with stuff in which they’re interested. Even for curiosity-driven research, that curiosity comes from somewhere. The challenge for us is to shine a light on the process transforming an applied problem into a research challenge that can give us a solution. It’s difficult to answer as far as specific metrics go, but I think what we need to start with is to clearly articulate what we are asking people to show evidence of, and engage in that conversation first.

KB: Building on that reframing of “purposeful research”, how do we go about identifying the problems that should inform the research we support? If there’s an existing company with a problem, it’s easy. What about broader societal goals that may not have an existing industry champion?

SL: Yes, the obvious case is where you’ve got a technical challenge. You know, if you’re Louis Pasteur, it’s fermentation in wine - it’s an economic disaster if they can’t sell their wine, and so that lends itself to an obvious set of challenges. The genius there is to broaden that into a wider set of applicable ideas.

What we need to be able to do here, and in fact what we’re required to do in order to give due response to the instructions we’ve been provided by our minister, is to begin with the three sectors that are the areas of focus within this economic model, and extract from that the research challenge components, because not everything should be treated as a research problem. Research is the right instrument to use when there is something we need to learn to better understand what to do. Under those circumstances, the application of scientific methods will get us closer to better, more informed choices. But there are many challenges where research is not the limiting factor. What’s preventing the improvement is something else: making the trade-off between competing benefits; devoting sufficient time or financial resources to something; or management or administrative competence. In those cases, calling for research is a diversion. There may be good things that come from that research, but it won’t be to improve the immediate problem at hand.

In an economic environment where the economy grows by industry participation, you could also look at through the lens that saving public investment means that you are reducing demand on public resources. But typically those are simply then drawn into other areas of public investment so it doesn’t necessarily produce cost savings. But if you accept the premise that you need industry to make use of what is produced, then the question of what should be the public investment versus what is a private investment opportunity becomes important. I think the appropriate area for public investment is where you have broad need, across a number of participants, where there is insufficient incentive for any one individual organization to pursue that particular work. You cannot expect an industry participant to take on that particular cost, because the spillover benefits will be to others. That makes a very strong argument for why the public sector ought to be involved at this stage, because it is going to provide broad sustained support to industry in that particular sector.

KB: Under the new strategic plan, does Research Nova Scotia take an active position on governance of the intellectual property coming out of the research you fund?

SL: One of the outcomes that we have been asked to produce, one of the seven that the minister’s given us, is to enhance commercialization of research, including retention and deployment of Nova Scotia IP in the province. So in the minister’s mind, the province is already thinking about this connection between who owns the IP and how it’s deployed in service of economic growth in the province as opposed to kind of the simpler question of what’s going to create jobs or whatever. So that was already part of the broad context in which Research Nova Scotia operates.

So as a result, we evaluated whether it made sense for us as an organization to take some sort of equity position or ownership of IP as a result of the research in which we were going to invest. We had a hard look, a really hard look, at a UK program which ran for about four or five years, where the model that they had was to invest public money in R&D programs executed by industry, but the public money retained the right to take ownership of any resulting IP if it wasn’t used within a particular timeframe. So if IP is created, if nothing’s done with it in that timeframe, then the public owns that IP. They went through three or four funding rounds and they were in an experimental and adaptive mode as we were discussing a few minutes ago. The reason we didn’t go down that path was that, in discussion with the developers of that program, they kept changing and trying to get closer and closer to delivery of a better program, and I realized that they had never actually exercised that opportunity because they didn’t know what to do with it. Just because they owned it wouldn’t necessarily then result in improved deployment of that IP in service of the ultimate outcome. Holding the patent is nothing, you have to actually then build it into something that then creates value.

I arrived at the conclusion that we need to stick with what we’re good at, and instead turn IP management into an area that an applicant for our research funding has to think through and shape and articulate to us. I think that this speaks to a difference between prescribing a set of rules that everyone has to work within, or developing an enabling framework with some flexibility that acknowledges that we might not know what the IP environment has to look like, but make clear to applicants that that part of our job is to ensure that there’s increased deployment of IP in Nova Scotia for its economic benefit. So we ask applicants to demonstrate to us how their approach is going to meet that outcome. Rather than us claiming IP or trying to manage it, applicants know they will need to develop a results distribution plan and that plan may be to share it broadly, it may be to hold it close, but it needs to be the right approach in context to deliver on the target outcome. It shifts the burden over to the applicant but it allows it to be flexible depending on what area they are working in. It’s going to be very different for deep tech versus process improvement. What IP actually means is going to be very different, and what’s patentable, and so on.

I have every interest in ensuring that the research we support creates commercializable value that is then picked up either by the performer or by whomever ownership passes to in order to mobilize it in service of society. But I’m very aware of my limitations and I do not assume I know how to do that better than them.

We need to be clear about what actually creates economic value. Is it the invention itself -- the output of R&D? Or is it the application of that invention to industrial activity, or cost-saving public service delivery? In my view, it is the latter. There is some economic value to R&D activity -- researchers are paid – but also there is typically greater opportunity to build economic activity adjacent to the R&D efforts. This is not an automatic process. Some clear-eyed assessment needs to be done about what core ingredients determine where R&D-driven economic activity will actually occur. If, in Nova Scotia, we have few to none of those ingredients (e.g. labour profile; domestic market; proximity to export; low carbon energy, etc.), then being the R&D centre is unlikely to result in the economic activity that comes with the application. I like to think about where we have a competitive advantage on applications (e.g. manufacturing or other forms of R&D deployment) that already exist or could be built, and focus R&D efforts there. These can be small niches (that’s inevitable with a population of 1M), but that’s where the opportunities lie. I like the concept of ‘conversion’: converting R&D results to application.You need to actually pay attention to not just your research strengths, but also the path through to converting that into value, in whatever form it may take.

KB: A central theme of your approach to all of this is “intention”, understanding and being tolerant to the fact that intention will not always translate to reality. Having researchers articulate that intention leaves lots of flexibility, but shifts the burden onto you in that your reporting systems have to reflect that flexibility. How are you thinking about making that open-ended reporting practical given the variety of approaches that you’re going to encounter?

SL: I should start by saying that like every publicly funded or taxpayer funded organization, there is an expectation from the people who fund us that we will be able to report convincingly that we are managing those funds responsibly and to further their aims. Nothing of what I’m saying is trying to shirk that sense of responsibility. But what I report has to be real. I don’t want to go to them and say that we can come up with a set of metrics that will allow you to track the dollar to the benefit when that is not a realistic expectation, and moreover I don’t think that’s really what they think is feasible either.

What’s really important is that, in everything we do, we understand what they’re trying to achieve, we take it seriously, and we adjust our processes and the people with whom we work and the type of research that we’re supporting, and we follow up wherever and however and in whatever creative ways we can to ensure that we’re able to report on outcomes and to take corrective measures. That is always going to involve a combination of two approaches. The first is a formalized approach in which we require certain reporting or provision of data that we can then assemble to a report. The easiest part of that is based on having a financial relationship with someone because the research process is underway, where they provide us with information and we give them funding in return–we can compel them. That’s easy.

The second part, and the only way this is going to actually work in the long term, though, is to maintain that reporting relationship over the much, much longer term, longer than we’ve yet been on this earth as an organization. How that’s going to happen precisely is difficult to prescribe at this point in time except that we have been doing this in a slightly different form for the first six years, and I can tell you that the way it actually happens is by cultivating a productive and trusting relationship with the researchers that you support over the long term, through which we build confidence that our role is not to provide oversight and pass judgment on their quality, but that our role is to enable their success.

“Success” here does not mean that I particularly care whether they’re going to get tenure or promotion, or publish in Nature or Cell, or whatever– those are the accoutrements of success. By success, I mean that the research that they are passionate about, and really want to see happen, is able to proceed through the system to its logical conclusion. That doesn’t mean it necessarily always works, and so cultivating that personal relationship means that they’re comfortable that when a challenge occurs, when supplier doesn’t have something, when something’s going to cost more than planned, when students who they thought were going to come can’t because the federal system changes and they can’t bring in a foreign student who was going to be the postdoc working on the project, or whatever the issue is, we want to be their first call. We want to be able to ask them how we can help overcome the issue. I think that’s how you build an understanding over the longer term of how productive work happens. I think mapping the journey is almost as important as how many people they hired, whether they filed a patent, and where that patent went. It’s going to be part of it but it can’t be all of it.

So yes, it shifts the burden to a more qualitative or narrative model, but I think ultimately that becomes more meaningful in terms of actually giving insight to the people who have given us the funds to further this kind of work as part of the many things they could spend their money on.

KB: You are leading the charge on a priority shift that I expect is going to come to the rest of the provinces, and probably the federal research agencies as well, in the near future. Having done this, do you have advice for your counterparts in other provinces that are asked to similarly reorient their research funding toward economic impact?

SL: You’re right that this is happening everywhere. Conversations we have with other provinces, at the federal level you talk about the Capstone Agency, mission-driven research, and dual-use tech etc. are all part of the conversation here, but in Australia, UK, the US, New Zealand, etc. everyone’s thinking about how granting agencies operate and and I think it’s important to consider that research ecosystem is very complex.

The shift is not just toward an economic focus. It is also in recognizing that there’s room for and a really powerful role for active coordination of research rather than just simply passively running a competition to see who applies, picking the best based on some research excellence metrics, and hoping for the best. I think that’s part of how research beauty happens, but it’s not the full thing. You need to round it out with a recognition that certain societal challenges need a different model. All the countries I just listed, and many of the different provinces, are faced with this challenge as well. I think maybe we had to confront it a little sooner in a real way than others and so I think, to revisit maybe some earlier thoughts and give emphasis, this idea of experimentation and adaptation is crucial.

The model that is has been used the dominant model that’s been used to identify research – say the peer-review system and all of that – has been relatively stable but that continues to undergo experimentation as well just in more subtle ways. We should expect, as we move toward a direct connection between economic activity and research outputs and a more coordinated model, to have a higher degree of adaptation or or experimentation early on.

The second observation is that for research funders and coordinators like us—which it to say, relatively small ones—are pulling macro-level levers. We are designing research funding programs, we’re engaging with universities as almost conceptual organizations, as groups of scholars, but this is going to work or not work by marrying those macro-level concerns with the micro reality. The micro reality involves, in academic settings, things like tenure and promotion systems, university senates, and the various decision-making structures within universities. In industry there’s the reality of supply chain, the cost of borrowing, and the competitive environment that’s changing. People are managing micro-forces and we’re pulling macro-levers. I think that understanding the dynamic of how those two aspects work together is going to be a determining factor in whether programs that we design and engage with are going to be successful or frustrating, or maybe how much they have to change.

I think that this has to be more than just the simple discussion around leveraging funds. A lot of the conversation revolves around the fact that when there is a new federal program, what’s going to be the percent of federal investment versus provincial investment. I think a more nuanced view of it is to consider what the federal role is, versus the provincial role. The outcome of a proper partnership-based discussion will be a funding model. But what matters when building R&D funding models that have both federal and provincial components is true reflection of what the different jurisdictions’ needs and capacities are. When I say ‘needs’, I mean that within a purposeful research program, those purposes are legitimately different. That is a good thing: let’s draw out where those purposes align, and where they don’t. The idea of ‘capacity’ is bigger than just financial: it includes things like decision-making structures and timeframes. This may express itself in a different matching formula when it comes to money but a partnership model actually looks at what each of the partners wishes to achieve and then finds something that in combination strengthens the position of the component parts.

I think that identifying and managing the spillover benefits and where they go really matters. This is bringing back the point that, at the provincial level, when we are directed explicitly to increase provincial economic growth and provincial economic productivity we really do need to pay attention to where that conversion into value is going to happen, and where R&D mastery converts to economic opportunity. We have to care about that. That’s not because we’re Philistines and don’t care about other issues in the world, it’s that we’re dealing with provincial dollars from provincial taxpayers. It’s the same person that’s a federal taxpayer, but in that sense they are contributing to a different set of interests than at the federal level.

The final thing, because it is something that we have really grappled with over the past eight months and I can’t imagine anyone doing this kind of work would fail to grapple with these two issues, is that coordination is really difficult. Coordination amongst funds that have different decision-making structures, timeframes, and mandates just takes time. Second, a lot of where research value can exist is in potential savings of public investment, say in healthcare. It’s really important to look at what those kinds of innovations can do to simply reduce demand in other areas. If you’re really focused on economic growth you actually have to look at what is going to grow the economy, not necessarily displace one expenditure for another. On the surface it seems somewhat simple to say “we are now devoted towards economic growth”. Understanding what economic growth actually looks like does matter even when considering what kind of research projects are best suited to achieving that kind of outcome.

KB: Is there anything I should have asked you but didn’t?

SL: I always ask people if they are optimistic. I’m personally optimistic even if I can be somewhat globally pessimistic. I think it’s a very challenging time, but I think that constraint is often a precursor to creativity. So this is, I think, the source of my optimism.

People ascribe this quote to either Ernest Rutherford or Winston Churchill but it goes something along the lines of “We have not got any money, so we have got to think”.

I’d love to have lots of research dollars to dispense. There’s lots of good work, there’s endless amounts of work and endless numbers of people who are doing great stuff, so I’m not saying that we need to be restricted in that regard but nevertheless it is to a certain extent those restrictions, which are not just financial but directional, and the reality of how the system works, that do offer us the opportunity to be creative and try something new.

I think there’s a reason it’s happening or it can happen here in Nova Scotia because we are in somewhat of a unique position. We don’t have the complexity that comes with the size of some of the other provinces. We are not nor will we ever have the budget of Ontario, Quebec, BC, or even Alberta. But we do have a substantial research history and research capacity that we can mobilize. We’re at that mesocosm level of having some of the elements that enable really high-quality, functional research to occur, but it is a population of about 1 million people that’s it, so we can connect on almost a personal direct basis to all the people that we’re working with. This is what gives me optimism: we can be and we should be and should be seen to be by others to be worth watching as an experiment to see what can be drawn into other jurisdictions as well. I’m an optimist in that regard.

Key Takeaways

Research with purpose

In the debate around Canada reorienting its research infrastructure toward the needs of society, I often come across the perception that funding research with the goal of socioeconomic impact is somehow incompatible with research conducted for pursuit of knowledge. As a result, the various words we use to describe the distinction between different research goals have picked up some baggage. Various commentators, including myself, tend to use simplistic labels that divide research into one of two categories variously called “fundamental”, “basic”, or “curiosity driven” and “applied” or “demand-driven”.

Stefan argues that this is too simplistic a categorization to be useful. Instead, he adopts a more nuanced framing around the idea of purposeful or intentional research.

“We looked at it from the perspective of identifying the “pull” and then deriving the research need from that, regardless of whether it is applied or fundamental by that set of definitions. […] For RNS, the discovery or the knowledge that’s gained through that research process has to be applicable beyond the immediate circumstances by people who weren’t involved in the initial investigation.”

In other words, it does not matter whether the research is fundamental or applied, what matters is that it responds to a clearly identified societal need. If the barrier to addressing that need is basic research, then that is what should be funded.

The challenge, then, is two-fold: identifying the need, and articulating clearly what kind of research needs to take place to address it.

If we consider the various existing research funding programs that focus on research commercialization (NSERC Alliance, I2I, and others), we find that they typically require an existing industry partner to co-fund the research, based on the idea that this signals clear private sector demand for the research. This is “demand-driven” research, which, while a subset of purposeful research, misses opportunities to address societal needs that do not yet have an industry champion. The RNS approach is more inclusive:

“You need not focus on evidence or expectation that whatever comes out of the lab is going to be applied, but instead look for indicators of how the problem within society that needs to be solved is influencing what goes on in the lab […] we need to completely change what we’re looking for, and focus on evidence of how the needs of society or industry shaped the research endeavor.”

In other words, funding agencies should look for clear evidence that their applicants are proposing research that is responsive to a need that they can clearly identify, and that the need has informed their research plan.

Acting on this involves a degree of flexibility on the part of the grant funders: while it is very easy to use the existence of an industry partner as evidence of societal need, we will have to do better if we are to build research funding systems that are more broadly responsive. Not all worthwhile problems have a company trying to solve them already. This has been a significant challenge for any public sector granting agency that seeks to fund research broadly, since it is impossible for public servants to be technical experts in all the areas of research they will encounter in grant applications. Stefan freely acknowledges the challenge:

“I have every interest in ensuring that the research we support creates commercializable value that is then picked up either by the performer or by whomever ownership passes to in order to mobilize it in service of society. But I’m very aware of my limitations and I do not assume I know how to do that better than [the researchers].”

Stefan’s reframing of what constitutes purposeful research provides a clear blueprint on which a more responsive research funding system should be based. The applicant is the technical expert. Instead of attempting to evaluate a proposal primarily on its technical merit while using industry demand as proxy for societal need, our research agencies should look for evidence that a proposal responds to a clearly identified need, and that the research plan is informed by a nuanced understanding of those needs and can articulate how the proposed research seeks to overcome the immediate challenge to addressing them.

Measuring success

Where research is concerned, intention will not always (or more accurately, will usually not) translate to the desired outcome. RNS has been handed a significant challenge in being asked to connect the research they fund to economic impact, since there are an enormous number of steps between the lab and societal impact that have nothing to do with the quality of the idea or the expertise of the researchers involved.

You can’t hold a researcher to account for a failure to realize economic return because there are so many aspects that are going to either frustrate or amplify creation of that economic value over which they have no control, but you can certainly expect that there is effort and intention behind what they do.”

The path from lab to market is often convoluted, and is only ever clear with the benefit of hindsight. This makes tracking outcomes extremely difficult, and leads to the typical proxy metrics that are used in most Canadian research programs: ;number of patents files, number of jobs created, etc. In my view, these should be considered input metrics rather than indications of output. Impact must be tracked on a much longer timeframe, one that is sensitive to what is reasonable in the context of different types of research. Stefan’s commentary supports this view:

“There was a study done not too long ago that looked at the timeframe between publication in a scientific journal, where the research has been “concluded”, and the application for a patent. Not the use of that patent, mind you, just applying it. That’s on the order of three to six years So, if you’re looking for evidence of economic benefit from the activity of research, you have to think about the timeframe of years. The more fundamental the research question, the longer that timeframe is. For basic research, we’re talking probably 20 years.”

Stefan notes that even when research eventually does achieve impact, the impact may not be felt in the same place the research was conducted.

“This is the challenge around R&D: you have a high degree of uncertainty in where those benefits are felt, but you have high certainty over where those costs originate. So you have what you could probably characterize as benefits that are diffuse; distant; hard to attribute to an individual supporter, and that are often disconnected from the original investor or research performer, but the costs have all the opposite characteristics.”

To respond to their mandate of provincial economic growth, RNS has to be clear-eyed about what it means to succeed, and they must be able to track the outcome of their projects over the long term, and that intention is just as important to outcome tracking as it is to identifying research worth funding:

“.[…] you need something that goes well beyond the timeframe of the research project itself […] and you need to focus as much on efforts and intent as on results.

Given the variety of research projects being funded and the diversity of ways in which research can respond to societal needs, there is no one-size-fits all approach to outcome tracking that will be effective. Stefan makes clear that rather than attempting to rely on a simplistic set of proxy metrics, their approach will be based on relationships:

“By success, I mean that the research that they are passionate about, and really want to see happen, is able to proceed through the system to its logical conclusion. That doesn't mean it necessarily always works, and so cultivating that personal relationship means that they're comfortable that when a challenge occurs […] we want to be their first call. We want to be able to ask them how we can help overcome the issue.”

Finally, iteration is key. Given the complexity, and the high probability that not all intention will translate to outcomes, the value of a tracking system is less about reporting on impact and justifying the money spent as it is about making sure that lessons are learned and incorporated into future funding calls.

“I think that the metrics you need to look at need to allow you to develop a sensitive understanding about why something worked, or why it didn’t work, so that you can apply those learnings to future investigations”

Outlook

Public funding for research has the potential to be enormously valuable to Canada. Just as not all research has an existing industry champion, not all societal needs can or will be addressed through private initiatives. In Stefan’s words:

“I think the appropriate area for public investment is where you have broad need, across a number of participants, where there is insufficient incentive for any one individual organization to pursue that particular work”

The commentary to this point makes clear that the debate should not be over what balance we should strike between funding fundamental or applied research, but rather that we should be thoughtful about ensuring that research is performed with intention and purpose. The words we choose to discuss policy challenges matter. I will be moving away from the idea of “demand-driven” research in my own writing and focusing on the more inclusive idea of purpose-driven and intentional research going forward.

Stefan’s closing remarks are particularly impactful. We are in a time of geopolitical upheaval, and it is more important than ever before that Canada learn how to translate its research excellence into real-world impact. Stefan comments that “constraint is often a precursor to creativity”, and it certainly feels to me that out public sector is waking up to the fact that doing more of the same thing we have been to date will not work, and that creativity is needed. This need for creativity in the face of constraint is the basis of my certainty that the change enacted by RNS will come to the rest of Canada in some form or another.

The strategic plan developed by RNS should be a central point of reference for all Canadian research funding programs that seek to achieve real-world impact from the research they fund. It reflects precisely what they ask of their researchers: a clear connection between the actions they are undertaking, and the needs of the society that they serve.

This will be the last CanInnovate post of 2025. Happy holidays!

A few months ago, the Conference Board of Canada (CBoC) published “Intellectual Property in Canada: Technology Specialization and Competitive Advantage”, the results of an impressive research process led by Zafer Sonmez. In this report, the CBoC team categorizes Canadian competitiveness in a variety of technology sectors through a careful examination of patenting trends, benchmarked against the world (and OECD) averages. Their results paint a bleak picture of Canadian competitiveness and IP ownership, but provide actionable insights for policy makers seeking to change this. I strongly encourage everyone to read their report, which is available freely at the link below.

Access the report

This wake-up call is perfectly timed, coinciding with a paper co-authored by myself and David Durand entitled “Intellectual Property is Economic and National Security”, in which we make the argument that sound IP management is critical for economic security and sovereignty. Where we provide the context that articulates why intellectual property is so important, the CBoC report highlights precisely where Canada can capitalize on existing strengths, and where we must improve.

In this post, I provide a brief overview of the report’s framework and methodology, and pull on a few threads that I think deserve specific attention. I will not cover the entire report, which contains a wealth of information and should be required reading for policy makers on the innovation file.

I would like to commend Mr. Sonmez and the CBoC team for an insightful and actionable contribution to an urgent debate about Canadian competitiveness and sovereignty.

A brief overview of methodology

The CBoC report is built on the basis of patent ownership data. Their conclusions revolve around two key metrics: the Relative Specialization Index (RSI) and the National Shift (NS). RSI, in a nutshell, measures the degree to which Canada holds a concentration of patents in a particular sector relative to that of the rest of the world, while the NS measures the pace of new patenting in that sector relative to the rest of the world. A high value of RSI indicates existing specialization, while a high value of NS indicates that Canada has a strong rate of invention in the sector, pointing to the possibility of emerging specialization. They use these metrics to categorize each sector considered as one of:

1. Leading inventive strength (RSI > 1, NS > 0), meaning that Canada has both existing specialization and a strong rate of invention;

2. Slipping inventive strength (RSI > 1, NS < 0), meaning that while Canada has specialization, the rest of the world is catching up;

3. Lagging inventive strength (RSI < 1, NS < 0), meaning that we are both behind and falling farther behind; and

4. Emerging inventive strength (RSI < 1, NS > 0), meaning that while Canada is currently behind, we are catching up the rest of the world.

The report further examines policy aimed at technological advancement and ranks each sector according to whether or not Canada’s inventive strength reflects stated policy priorities (spoiler: it mostly doesn’t).

Strengths, weaknesses, opportunities, and threats

It is interesting to note the mismatch between the typical language around Canadian specialization as reported in the various national innovation strategies and the realities exposed by this report. For example, Canada is generally regarded as having a strong biotech sector, and while we do indeed have specialization here, the report makes clear that we are resting on our laurels and in danger of losing this distinction.

Nanotechnology stands out among the surprising opportunities in the report. While we lack strong specialization, it appears that we are on track to change this. This is a trend on which we should double down.

Probably the most striking result was the near-complete misalignment in technology sectors relating to agriculture. As was highlighted in my previous interview with Mark Olson, Canada has the potential to be the world’s bread basket. We have unmatched potential in agriculture, with abundant natural reserves of nearly everything necessary while being one of the few countries on the planet whose capacity for food production will not be negatively impacted by a warming world. Canada could project power through food export over the next several decades if we chose to use our myriad natural advantages.

At FORPIQ 2025, a panel discussion on Canada’s potential as an agricultural powerhouse highlighted that the second largest food exporter in the world (after the US) is the Netherlands. This is based entirely on highly efficient and innovative agricultural practices, given that their available arable land area is miniscule compared to Canada’s. Their examples demonstrates just how much room we have to improve. Our failure to secure competitive advantage in this area represents an enormous wasted opportunity for sovereignty-focused economic development.

One of the most concerning issues highlighted in the report relates to the distribution of patent ownership. CBoC shows that Canada lags far behind where organizations with large patent portfolios are concerned - but patent ownership tells only a part of the story. Where access to or control over patents is granted through licenses instead of assignment, it will not show up in the ownership data. In other words, the reality is even worse than what the report shows.

The fact that licensing details cannot be extracted from public data points to a systemic failure to collect that data rather than any weakness of the report. Rather, this highlights the importance of ensuring that data is collected on control over and access to Canadian IP, especially in cases where the IP is the result of publicly funded research. The recent submission by the SAIL team to the House of Common Standing Committee on Science and Research, as well as the guidance we have previously built around the Simple Agreement for Innovation Licensing (SAIL), give some more detail on the kind of data needed to address this.

What’s next

While it is to be expected that Canada lacks incumbent competitive advantage in many sectors, the number of sectors that are deemed national priorities in which we lack even emerging inventive strength should be a wakeup call for policy makers that current fragmented efforts are not effective.

This is not just a matter of economic competitiveness, but also one of economic security. Priority sectors in that Canada funds heavily but in which we lack emerging inventive strength in particular should be reviewed carefully, as these sectors are likely to be leaking IP. The CBoC report and our paper on the connection between IP and economic security combine to form a roadmap to address these challenges: Allocate funding to the early stages of research commercialization in sectors of national interest in which Canada lacks competitive advantage, and use slipping and lagging inventive strength as bellwethers for weaknesses in Canada’s approach to IP management.

Above all, recognize and prioritize investment in Canada’s natural strengths. If Canada is to secure its economic and technological sovereignty, we will need to be strategic in our approach. We are not big enough to be leaders in every sector, but we are well positioned to dominate a few. This report shows clearly how our inventive resources can be allocated to achieve maximal impact.

In our work connecting IP to economic security, one of our core recommendations was to create an open-source intelligence agency that continuously monitors emerging technology trends through analysis of patents and research funding. This CBoC report shows us just how valuable that capability would be.

Last week I appeared as as witness before the House of Commons Standing Committee on Science and Research (SRSR) to discuss ways to overcome Canada’s challenges turning research into economic impact.

I was encouraged to note that I was not the only voice in the room calling for risk-tolerant capital to be deployed in support of research commercialization. Other notable points made by my colleagues (in particular Dr. Gerry Wright and Kevin Dahl from ElevateIP Alberta, with whom I am in full agreement), relate to a pressing need for

1. funding specific to early-stage startups commercializing research;

2. funding for IP creation, protection, retention, and strategy;

3. IP-based valuation tools and IP-backed lending mechanisms; and

4. moving away from input metrics like research spending as indicators of impact and toward actually measuring the resulting outputs.

My presentation is shown in the excerpt below, and you can watch the full proceedings at the link underneath.

Watch full session on ParlVu

Transcript of Opening Remarks

Thank you Honourable Chair Zahid, Vice-chairs and Committee members for the invitation to speak to you on commercializing innovation emerging from publicly funded research at Canadian universities.

To briefly introduce myself, I am a physicist and the Entrepreneur in Residence at the Faculty of Science at the University of Ottawa. I was the CEO of Northern Nanopore Instruments, a nanotechnology startup built from my PhD research that we sold in Fall 2023. I am now the author of the CanInnovate blog on innovation policy and, alongside TJ Misra and David Durand, co-author of the Simple Agreement for Innovation Licensing (SAIL) and co-founder of the SAIL Initiative, through which we aim to streamline how new technologies get from lab to market.

The value of research follows a power law distribution: a small minority of new technologies eventually create most of the economic value, but it is impossible to reliably predict which ones will succeed or be valuable when the research is transferred to the private sector. Because of this impossibility, it is more important to make sure that we do not miss the valuable minority than it is to make sure that every attempt is successful. It follows that the most effective strategy for commercializing research successfully is to invest relatively small amounts of capital early in almost everything. Mostly, Canada does not do this.

Attempts to commercialize research have failure rates above 90%, but the successes create more than enough value to offset the cost of the failures. Countries that “invest well” have three things in common, namely, the public sector:

1. funds innovative startups before they have revenues,

2. usually favours new startups over existing companies as vehicles of innovation, and

3. is willing to let previously-funded projects fail when necessary.

This approach to funding serves to derisk innovative companies to the point where they can attract private sector investment, and ensures that valuable technologies do not slip through the cracks.

Ironically, systemic risk intolerance in the Canadian public sector mostly prevents it from funding pre-revenue companies in this way, and Canadian VCs cannot address the gap because of the long timelines involved. We must reframe how we evaluate and manage risk. We must understand that most investments will fail, and that this is acceptable as long as the combined return over time is positive. So while it may be true to say that our public sector has become too risk averse, I argue that “investing in almost everything” is in fact less risky (or at least more likely to produce a positive outcome).

Another element common to effective commercialization elsewhere is harmonized innovation policy. The United States has the Bayh-Dole Act, for example, that guides how universities transfer technologies to the private sector. In Canada, we lack even an attempt at national coordination. Senator Colin Deacon’s office recently found 134 different innovation funding programs at the federal level alone, and the tri-council agencies provide no top-down guidance on how universities should manage research IP. As a result, every research institution has a different IP policy, licensing negotiations with most research institutions are slow, no two licences are alike, and there is no standard for data collection on licensing or outcomes.

To address these challenges, my colleagues and I developed SAIL. SAIL is a licence framework designed to support harmonized and streamlined Canadian tech transfer from research institutions to startups. After consulting with a national community of innovation stakeholders, we designed SAIL based on 6 axioms of tech transfer specific to Canada’s unique challenges. It asks universities to play the role of “first investor” in research commercialization and rewards them with a pre-defined amount of convertible debt in exchange. The framework can be easily amended (with legal advice) to support a variety of startups more efficiently and effectively than building a new licence from scratch every time.

We are also working to adapt a risk-tolerant funding mechanism that has been highly successful in the UK, where it created an estimated $7 of economic value for every $1 of input. This model uses venture philanthropy delivered through a public-private partnership, combining public funds with private donations and university contributions to create a charitable investment fund that reinvests all returns to ensure that it is self-sustaining. Variations on this model have been implemented at a handful of Canadian research institutions, most notably the UCeed fund in Calgary, and we propose to implement it at the national level.

Policies to promote and grow private sector investment in research can only be effective if we increase the pool of investable, innovative companies. To do that, we must first build a better bridge from lab to market. I recommend that Canada:

1. embrace strategic risk-taking: deploy public funding toward pre-revenue startups commercializing Canadian research by trialling venture philanthropy delivered as a public-private partnership with a national scope, and

2. nationally harmonize a Canada-first approach to management of the IP arising from publicly funded research.

Thank you distinguished members of this Committee for your time. I look forward to your questions.

About a year ago, I was asked by the Canadian Council of Academies (CCA) to write an evidence synthesis paper on the state of deep tech in Canada. The paper that resulted, which I submitted to CCA in February 2025, is 1 of 8 that informed a report entitled “The State of Science, Technology, and Innovation in Canada 2025”, commissioned from CCA by Innovation, Science and Economic Development Canada (ISED). The completed report, as well as the paper I contributed entitled “Challenges and opportunities for Canadian deep tech commercialization”, are now available publicly.

In this post, I provide a brief overview of my contribution, since it synthesizes in one place many of the key themes that have informed by writing on CanInnovate since I started this project. The other papers commissioned to support the CCA report are also well worth a read.

Many thanks to CCA for the opportunity to write this paper, and to the peer reviewers whose commentary resulted in a much stronger contribution.

What is Deep Tech?

In the course of building my own deep tech company and in all my subsequent writing, I realized that the phrase “deep tech” means something different to everyone who uses it. For this reason, someone watching carefully might have noticed that I have moved away from using it in my own writing, because it can easily be the basis for misunderstanding unless one takes the time to carefully define it every time. This became the first task of the paper.

After careful review, I adopted the definition presented in an earlier review by Romasanta et al., who come to the conclusion that deep tech is:

“Early-stage technologies based on scientific or engineering advances, requiring long development times, systemic integration, and sophisticated knowledge to create downstream offerings with the potential to address grand societal challenges”

There is a lot of complexity buried in this definition, which the paper attempts to unpack and use to inform a detailed discussion of the challenges and opportunities for Canadian deep tech commercialization. I will be using this definition in my own writing going forward wherever deep tech comes up.

Challenges and Opportunities

In the paper, I focus on articulating challenges with deep tech commercialization that revolve around a few key themes. The first is IP governance and harmonized innovation policy. I compare the approaches used by the United States via the SBIR program, the approach used by the Israeli Innovation Authority, and Canada. As with previous articles that touch on these themes, I stress that while Canada has a lot to learn from other jurisdictions, it is not possible to directly import policy frameworks that work elsewhere—we must adapt them to the Canadian context first. As I have written in other contexts, Canada must harmonize its approach to innovation generally, and to intellectual property governance in particular, if we hope to become effective at commercializing deep tech.

The second main theme of risk tolerance winds through all the commentary in the paper. Risk-tolerant public funding, delivered to startups and small companies commercializing deep tech, is a common theme of innovative ecosystems that are effective at extracting economic value from deep tech, and lies at the core of Canada’s problems with commercialization deep tech and research. I draw on the example of the American Small Business Innovation Research (SBIR) program and compare it to the recently cut Canadian Innovative Solutions Canada, arguing that Canada will need to rely more heavily on active coordination between market demand for deep tech and the research that generates it than our southern neighbours.

The third main theme is data, and Canada’s systemic failure to collect metrics of performance that are useful for policy reform. In the paper, and drawing on work by others, I suggest that we must improve our data collection practices in any new initiatives aimed at improving our situation, acknowledging that intervention and impact may be separated by many years where deep tech is concerned. I also point out that university-focused investment funds that deploy patient capital are uniquely positioned to collect that data, provided that the approach can be standardized, and that long-term tracking of control over and access to research IP assets should be prioritized as an input to future policy updates and mandated as a condition of research IP licensing.

Finally, I spend some time debunking the idea that Canada’s challenges with deep tech commercialization are cultural.

The report concludes with reviews of three sectors with overlap in deep tech: quantum technologies, cleantech, and artificial intelligence (AI). The lessons that I present from review of related literature strongly support the ideas that

1. Canada’s quantum advantage provides a blueprint for deep tech dominance that can be replicated,

2. Greater coordination is needed between the various moving parts of the innovation ecosystem in areas (like cleantech) in which regulatory complexity is high, and that

3. Canada’s approach to AI is a cautionary tale of what happens when we fail to act sufficiently quickly to secure a deep tech advantage, but that all hope is not lost.

As I note in the conclusion of the paper:

“Canada has all the raw ingredients it needs to be a globally relevant force for deep tech commercialization.”

The CCA report is a strong contribution to the debate around how best to combine these ingredients, and I am grateful to have been given the opportunity to contribute. This paper represents a synthesis of many of the themes and ideas that have informed the articles on CanInnovate to date. I hope you enjoy reading it as much as I enjoyed writing it.

This week I interviewed Bruno Couillard, the CEO of Crypto4A, an engineer who has spent a decades-long career building the hardware security solutions that have underpinned trust on the internet for decades. Through Crypto4A, he seeks to do the same for a world where quantum computing is a reality and where legacy cryptographic solutions are no longer an option.

My goal in this interview was to better understand Canada’s contribution to this space and the opportunities before us as we reorient our economy toward security, sovereignty, and defense. Bruno’s insights touch on the foundations of what digital sovereignty means practically and how security and trust on the internet will evolve as we move toward a world in which quantum computing is a reality.

Bruno also makes clear that Canada has made globally leading contributions to cybersecurity that it has mostly failed to benefit from economically to date, and that we have a rare second chance to get it right as quantum computing becomes practical and the world shifts its priorities toward digital sovereignty.

Your email client will probably truncate this post. My key takeaways are presented at the end, so be sure to read the web version if you want to get the whole story. Many thanks to Bruno for sharing his experience and insight.

Interviewer’s note: Bruno Couillard and the Crypto4A team approved the final version of the section entitled “Interview with Bruno Couillard” and had editorial input on that section, with the option to rephrase and expand on the ideas discussed in the interview without changing or removing any intended meaning. The key takeaways presented at the end are my own commentary, and do not necessarily represent the views of Crypto4A or its employees.

Interview with Bruno Couillard

KB: How’d you find your way into quantum cryptography?

BC: My journey into cryptography and key management began in 1988. I had just finished my military officer training at the Royal Military College, and I was assigned to a fascinating project in Ottawa called the Integrated Data Network (IDN). It was essentially a Canadian Forces–wide internet, connecting all bases across Canada and even the two we had in Germany at the time.

The IDN relied on a mix of link and end-to-end encryption, securing both classified and protected information. Fresh out of engineering school, I was eager to contribute, and I got assigned to the cryptography and key management side of things. Back then, these concepts weren’t taught in school you had to learn them on your own. I dove in, and it was love at first sight. That set the course for my career: cryptography, key management, and public key infrastructure have been my focus for 37 years now.

From IDN, I moved to the Communications Security Establishment (CSE) in 1992. Then, just as the Internet was taking off in 1994, I left to co-found Chrysalis ITS in Ottawa. As CTO and co-founder, I transitioned from theory into practice designing and evolving hardware products. That’s where I created the Luna HSM, which became a leading Hardware Security Module worldwide. I also helped shape standards like PKCS#11 and contributed to best practices around root key ceremonies. After almost a decade, I left Chrysalis in 2003.

Editor’s note: A Hardware Security Module (HSM) is a dedicated physical device designed to securely generate, store, and manage cryptographic keys. HSMs perform operations such as encryption, decryption, digital signing, and key generation within a tamper-resistant environment. They are foundational to the security of systems that require strong encryption and regulatory compliance such as banking, cloud operations, and secure communications, and are key elements of public key infrastructure (PKI).

At that point, I wanted to recharge and spend more time with my family. My wife and I have three sons who were busy with school, sports, and music. I rebranded myself as a consultant and returned to projects with DND and CSE. It felt like going back to my roots, but with new perspective: instead of focusing on commercial HSMs, I was working on high-grade security systems.

Around 2003, one of the major efforts was modernizing cryptography and key management across allied security systems. The idea was to migrate algorithms quickly enough to withstand advances in general-purpose computing. I was heavily involved in Suite A and Suite B crypto right from their inception.

Then, in 2009, the NSA published a pivotal statement: it wasn’t general-purpose computers we needed to worry about anymore, it was quantum computers. Practically overnight, post-quantum cryptography (PQC) became the main focus. That announcement grabbed my attention. I began watching closely, wondering if anyone in the commercial HSM market was preparing for this challenge. To my surprise, no one was.

That’s when I started thinking about building a new HSM from scratch something designed from the ground up to be quantum-safe and ready for the fast-evolving digital economy. Together with some long-time colleagues from Chrysalis, we decided to take the leap. And that’s how Crypto4A was born.

From day one, our vision was clear: Crypto4A would create the HSM built to survive and thrive in a quantum era. These devices matter because they sit at the very center of the digital economy, which now represents more than a third of the global economy. This all traces back to 1995, when Netscape introduced SSL for secure browsing. Behind that were authentication certificates issued by Entrust or Verisign, and guarding those authorities was the Luna HSM. That triad, SSL, PKI, and HSM remains the foundation of digital trust today.

Over the past 30 years, we’ve only layered more services onto that same foundation: banking, government, identity, communications, everything. And behind it all, somewhere in the background, an HSM protects the root keys of trust. Having been part of the team that created the first successful generation of HSMs, I felt it was our responsibility to build the next generation this time quantum safe.

Life has given me incredible opportunities to be at the right place, right time, with the right people, participating in one of humanity’s greatest transformations: digitizing a third of the global economy in just three decades. Now, we’re entering the next major transitions; AI today, quantum tomorrow. But none of it will be possible unless the foundations of digital trust we laid down in 1995 are made quantum-safe and continue to hold strong for decades to come.

KB: Where are Canada and the rest of the allies in terms of replacing their hardware stack to make it quantum resistant?

BC: If you had asked me this question six to eight months ago, I would have said the United States was clearly leading the world in this push, with Europe and Canada following behind.

But today, I’d say Canada is really stepping up to the plate and starting to take a much more significant leadership role. What’s interesting is that this is happening at a critical moment for the digital economy. Two major themes, trust and sovereignty, have now joined the conversation alongside quantum readiness.

If we’re going to build the capabilities and technology stack for the future, its foundation, cryptography, must be quantum safe. That means protecting the root of trust while also ensuring sovereignty of its operations. Put simply, there is no digital sovereignty without cryptographic sovereignty. And this is exactly where HSMs come in, and where Canada has a real opportunity to lead.

On the political front, our leaders have been making bold statements over the past few months. We’re hearing strong, visible messages from the highest levels of government about the sovereignty of our information and the urgency of preparing for the quantum era. Even at the recent G7 meeting in Canada, quantum realities were flagged as a concern on the global agenda. Data sovereignty is no longer a niche issue; it’s now front and center here at home and around the world.

Of course, the real test will be whether these political commitments can make their way through the machinery of government to create meaningful action. That’s never easy. Cryptography underpins so much of our world that it’s almost easier to list the things that don’t use it than the ones that do.

The good news is that Canada has an incredible foundation to build on. We’ve been investing in cryptography and key management since the early 1990s, when PKI and HSM technologies first took root here. That’s when both Entrust and Chrysalis-ITS were born out of the Nortel/BNR era. Those companies became the backbone of the modern digital economy, and their DNA is still very present in Ottawa today.

The Luna HSM design team, now part of Thales, still works out of Ottawa South. Entrust’s PKI team is still in Kanata. Crypto4A itself was founded by former Chrysalis-ITS and Entrust engineers. Taken together, Ottawa, and by extension Canada, likely represents the highest concentration of cryptographic and key management expertise anywhere in the world.

For some reason, we became very good at this discipline. And, true to Canadian form, we don’t brag about it. But we should. Because the reality is: Canada owns some of the most critical pieces of this global puzzle. And given the scale and complexity of what lies ahead, I’d much rather be tackling this challenge here in Ottawa, Canada, than anywhere else on the planet.

KB: Beyond encryption breaking, what’s the business case for quantum cryptography? What’s real and what’s hype in the current thinking around capabilities relating to quantum cryptography?

BC: Before I dive in, let me refine the question a little. The real capability we’re worried about isn’t just “quantum cryptography” in the broad sense, it’s the arrival of quantum computers powerful enough to run Shor’s algorithm. Once that becomes possible, whoever controls that capability will be able to break virtually all the classical public key cryptography we rely on today. That’s why there’s a global race underway to migrate as quickly as possible to post-quantum cryptography, or PQC.

Now, when people hear “quantum cryptography,” they often confuse two very different things. What I just described is PQC developing new, quantum-resistant algorithms to replace the ones we use today. But there’s also a field properly called quantum cryptography, which uses the principles of quantum mechanics, things like entanglement and superposition to achieve new forms of security.

One of the best-known examples is Quantum Key Distribution, or QKD. It allows cryptographic keys to be generated at two different locations in such a way that both parties can be confident the key is identical, and that it hasn’t and in fact can’t be copied or intercepted anywhere else in the universe. This concept was first introduced in 1984 by Charles Bennett and Gilles Brassard in what became known as the BB84 protocol. Since then, several variations have been developed, but they all rely on the same unbreakable quantum mechanics principles.

We’ve already seen some remarkable demonstrations. China, for instance, has deployed advanced QKD networks and even tested satellite-based systems where secure keys were exchanged between Earth-based stations and satellites in orbit. These are fascinating achievements and hint at some powerful niche applications in the future.

That said, it’s important to keep perspective. QKD isn’t a wholesale replacement for the cryptography we use today. It addresses certain use cases very well, but it won’t solve the broader challenge. Most of our digital systems from banking and government to cloud and communications will continue to depend on PQC. In practice, QKD should be seen as a complement to PQC, not a substitute.

So, to your question of what’s real and what’s hype: the urgent and unavoidable business case is defensive replacing our current cryptography with PQC before Shor-capable quantum computers arrive. QKD is real as well, but more specialized, and won’t replace everything. The hype comes when people conflate the two or imagine QKD is going to single-handedly secure the digital world. PQC is the foundation we must address first, and QKD is an exciting complement that will play a role in specific scenarios.

KB: Can you speak to the limitations that prevent QKD being a replacement for more traditional cryptography methods?

BC: The limitation with QKD is really about practicality and scope. For example, if you and I wanted to set up a QKD-secured session between our two locations, we’d need a dedicated fiber optic cable connecting us. If the distance was too long, we’d have to rely on multiple QKD segments or even use a satellite-based system. That works well for “real-time” sessions securing data in transit, but it doesn’t solve long-term security needs.

Take something as simple as buying a house. If I sign a digital contract today, that signature has to remain valid not just tomorrow, but 20 or 50 years from now, so I can still prove I own my home. That requires a digital stamp that will stand the test of time. QKD can’t do that. Only PQC algorithms can provide the kind of durable digital signatures needed for those use cases.

The same applies to countless everyday examples. If you buy a smart car, the manufacturer will send firmware updates for years. Each update must be digitally signed to guarantee authenticity and integrity. In the quantum era, PQC algorithms will be the only way to achieve that. The same holds true for critical medical devices like pacemakers or insulin pumps, where firmware updates must be trusted without question.

That’s why there’s such urgency around migrating these systems now. Once a satellite is launched, or a medical device is implanted, or a long-term legal contract is signed, the cryptographic signature tied to it has to remain secure for decades. Those are “long-lived signatures,” and they absolutely must be quantum safe.

If you look at the landscape, there are thousands upon thousands of use cases where we’ll need to authenticate, verify, and preserve the integrity of digital artifacts in our lives. All of those still exist in the digital world of ones and zeros. We haven’t moved into a pure world of photons and light yet and we likely won’t in our lifetimes.

So, for now, and for many years to come, our security will continue to depend on PQC. QKD is exciting, but its use cases are in the dozens, not the thousands. It’s worth experimenting with, but it’s not “the next security solution.” It’s a complement. The heavy lifting will remain with digital cryptography zeros and ones for a long time.

KB: “Quantum” as an umbrella term has entered pop culture and often means different things to different people. What are some common misunderstandings and misconceptions that you encounter?

BC: Early on, there was a lot of hype around quantum. People talked about it as if it were a panacea that it would solve every problem and usher in a perfect future where we all lived on a cloud. The reality is much different. Quantum technologies are extremely hard, very expensive, and although they will eventually be amazing for many applications, we are still far from realizing all their promises. One day, yes, we’ll use quantum computers as naturally as we use classical computers now. But today, we’re still in the early stages.

This hype phase has a downside: customers get bombarded with messages that range from the realistic to the completely ridiculous. That noise makes it difficult for those of us focused on a secure transition to the quantum era to cut through especially when it comes to post-quantum cryptography. Too often, people think the solution is as simple as deploying quantum key distribution everywhere.

Another challenge is language. Many of the terms are not yet fully standardized, so people use them interchangeably even when they mean very different things. Fortunately, if you read Canada’s National Quantum Strategy, or other national strategies, you see a clearer framework emerging for the different “buckets” of quantum technology.

The first bucket is quantum computing. These are machines that replace today’s ones and zeros with photons or other quantum particles, and that can run true quantum algorithms. One of the most powerful, and frightening, of these algorithms is Shor’s algorithm. Discovered in 1994, it can break every classical public key cryptosystem: RSA, DSA, ECDSA, Diffie-Hellman, and more. This is what launched the global race to build such machines, with billions invested so far. Roughly half of that investment has come from China, and it would be naïve to think their priority is only better materials science or drug design. The more immediate goal could be breaking our current cryptography. If that happens, the consequences would be catastrophic: our banking system would fail, our digital economy would collapse, and society itself would grind to a halt because the Internet could no longer be trusted. Every sector of our lives, from mining to farming, already depends on secure digital connections. Without trust in the Internet, modern society doesn’t function.

The second bucket is quantum sensing, which I find fascinating. By harnessing quantum effects, you can detect incredibly faint signals: mapping mineral deposits deep underground, exploring the oceans, reading subtle electrical signals in the human brain, even probing signals from the far reaches of the universe. Compared to these advances, our classical sensors will look primitive. Of all the quantum technologies, I think sensing may have some of the most profoundly positive impacts.

The third is quantum communication, and this is where confusion tends to multiply. It includes things like quantum key distribution, which we’ve already discussed, as well as quantum teleportation, the idea of transmitting particles across distances to create new forms of communication. Another element is quantum random number generation, where we harness inherently unpredictable quantum processes, like radioactive decay or photon behavior, to create true randomness.

Finally, there is post-quantum cryptography. Strictly speaking, PQC is not “quantum” in that, it is not strictly speaking based on quantum mechanical principals at its core. Instead, it uses advanced mathematical algorithms designed to withstand both classical and quantum attacks. PQC exists for one reason: to ensure our digital infrastructure remains secure in a quantum world. At Crypto4A, we’ve built a new generation of cybersecurity products around this principle. Our solutions still operate with ones and zeros, but they use the new algorithms that NIST has been standardizing after an eight-year global competition. Three were standardized in August 2024 and more will follow in the coming years. From now on, it’s not enough for systems to be quantum-safe they must also be crypto-agile, able to adapt as algorithms evolve.

This is where misconceptions really take hold. Some vendors will say that because they have a quantum random number generator, their products are “quantum ready.” Others claim QKD or teleportation will solve every security challenge. Some even suggest they can magically make a classical device quantum-safe with a quick patch. None of that is true. To be truly quantum-ready, you have to redesign your entire stack to survive in a world where quantum computers exist. Every element has to be tested against that future reality. The litmus test is simple: imagine your design in the year 2040. Quantum computers are everywhere. Could a hacker in that world break your system? If the answer is yes, then you’re not quantum ready.

KB: Taking that one step further, what should governments and businesses be doing right now? How long do they have to get it done?

BC: We already solved this kind of problem once before. Thirty years ago, November 1995, we figured out how to make the Internet a trusted environment for business. That was when Netscape introduced SSL, completing what I call the “digital trust foundation.” SSL secured communications between browsers and servers, PKI technologies from Verisign and Entrust allowed Certificate Authorities to issue SSL certificates, and the Luna HSM that we had developed at Chrysalis-ITS here in Ottawa, underpinned the cryptography at the core. Those three pieces, SSL, PKI, and HSM, are still the same foundation the Internet rests on today.

To get to a quantum-safe world, we need to repeat that process. First, deploy quantum-safe HSMs. Then, make PKI quantum safe. And finally, update Internet standards and protocols so they can use quantum-safe certificates. You start at the foundation and work your way up the stack. Whether you’re a bank, a government, a manufacturer, a transport system, or an energy grid operator, the first step is the same: begin deploying quantum-safe HSMs.

At Crypto4A, we’ve built these next-generation HSMs. They’re quantum-safe by design, highly specialized machines that we can produce at scale. But like any physical component, supply is not infinite. Think back to COVID, when everyone suddenly needed masks and gloves, the lineups were endless, and prices spiked. The same will happen if organizations wait until the moment a breakthrough in quantum computing arrives. If you start now, the migration can be steady and manageable. If you wait, it will be chaotic and costly.

There’s no downside to moving early. I always recommend that governments, banks, and anyone providing trusted services get started right away. Canada has a chance to turn this into a national advantage. We’ve always punched above our weight in cybersecurity. If we make quantum-safe migration a major national project, Canada could cement itself as a global leader for decades to come.

And this isn’t just about technology. The digital economy, the intangible economy, is already outpacing the tangible one. In that reality, leading the world in securing cyberspace is the equivalent of having the strongest defense forces. It’s a matter of national strength and sovereignty.

KB: In a military context there is a need to build systems that allow for interoperability with allies, which points to the need for standards that are agreed internationally. Does Canada have a role to play in defining those standards?

BC: I’d say Canada is already very involved in shaping these standards. Over the past 30 years, much of cyberspace has been built through the work of international standards bodies groups of volunteers who dedicate their time and expertise to advancing the protocols we all depend on. One of the most important of these is the Internet Engineering Task Force, or IETF, which was founded more than 60 years ago and continues to evolve the core protocols of the Internet today.

Crypto4A actively participates in this work. We’ve contributed to efforts at the IETF, and we’ve been deeply involved in the post-quantum cryptography standardization process led by NIST. Our team has consistently pushed the envelope on quantum-safe design, because our mission is to ensure cyberspace remains secure and trusted for generations to come.

We’re also proud to be the leading HSM vendor in the quantum migration project run by the National Cybersecurity Center of Excellence (NCCoE), a lab under NIST in Maryland. That project brings together the world’s top cybersecurity companies to demonstrate interoperability of quantum-safe cryptography and related standards. Since its launch in 2022, we’ve led the HSM interoperability testing and have consistently delivered fully functional, standards-compliant products some of which are already deployed in live operational systems today.

And it’s not just us. The Government of Canada itself has a strong presence in these efforts, with experts participating in NIST’s standardization process, the evolution of IETF protocols, and other international bodies. Taken together, Canada’s contributions are significant. We’re not just observers we’re helping shape the very standards that will define how the world transitions to a quantum-safe future.

KB: Following the Luna HSM over the years, it has been the target of 4 acquisitions, the first of which caused Canada to lose control of this IP. It is still in service now, decades later, as part of the Thales Group IP portfolio. Tell us the story of what led to the original decision to sell the IP. What needs to happen differently this time around if we want Canada to lead in the post-quantum HSM space?

BC: After coming out of a military and government career, I wanted to build something to put into practice all the theory I had developed on how to ensure trust in the digital world. I co-founded Chrysalis, focusing on the technical side of things while my co-founder ran the business.

From 1994-1998 we designed the Luna HSM, and by 1998 we were selling it to the US, Canada, a few European countries, and Japan. A lot of our sales were alongside Entrust sales or directly to Verisign, forming the foundations of public key encryption (PKI). In 1998, the dotcom bubble was well underway, though, and chip-design was all investors wanted to hear about. Our investors decided that we needed to design a chip. We tried and failed to find a contractor that could design it, and eventually hired some very expensive in-house expertise to do it. We almost did, but in 2001, the board decided to pull the plug on all the semiconductor work. My co-founder was let go on August 15, 2001, along with 100 other people, and I was told to stay on as CTO and to go back to our roots with the Luna HSM and a number of other projects. September 11 happened a few weeks later, and the company almost disappeared. In 2003, after two years of intense efforts, we’d managed to turn things around with the introduction of the world’s first network-attached LunaHSM with very promising sales results. At that point, I decided to refocus my energy on my young family with my wife, left Chrysalis and transitioned into a consulting role with the Canadian Government. Shortly thereafter, Chrysalis was sold at a bargain price, basically just to acquire the IP.

That Luna HSM IP went on to become the global standard for its purpose and is still in use today, but it was lost to the Canadian ecosystem in 2003. It has since been sold several more times - from Chrysalis it went to Rainbow Technologies, then SafeNet, then Gemalto, and most recently to Thales. Given the value that IP has created in the world since it was first sold, I think that a Chrysalis that had stayed fully focused on the HSM instead of getting side tracked by semiconductor hype might still be alive and Canadian today. If there was a mistake made, aside from the semiconductor mess, it was simply not recognizing the value of the IP that we had and being relatively junior to the process of building a company.

Crypto4A is a second chance to do the same thing for a post-quantum world what LunaHSM did for the early internet: ensure the future strength, security, and trust of the internet in the upcoming quantum era. Someone is going to do it, because otherwise the internet stops working.

This is especially critical as we start to talk about data sovereignty. The only way you enforce security capabilities is by controlling the keys used for the cryptographic processes that secure your data. Without that control, all bets are off. You cannot ever allow data to be sent abroad and let someone else control the keys: no key, no sovereignty.

I believe what we are building is the world’s best technology for its purpose and that it will have the same longevity as the Luna HSM. I can assure you that I know far more today than I did when designing the LunaHSM - we will not make the same mistakes twice.

KB: What should I have asked you but didn’t?

BC: One of the biggest challenges right now is clarity, separating truth from noise. There’s so much misinformation, hype, and fearmongering out there that it becomes very difficult for people outside our “inner circle” to focus on what really matters.

Decision-makers today are bombarded with messages and marketing claims, many of them contradictory. That’s why your role as someone who listens to these different voices and distills them into something understandable and neutral for policymakers, lawmakers, and planners is so critical. You may not be an expert in quantum technologies or cryptographic algorithms, but you still have to make sense of it all and present it in a way that drives sound decisions. That’s not easy.

It’s something I think about a lot in my own organization. We recently brought in our first marketing lead, and one of my requirements was that they not come from a cryptography or key management background. I didn’t want someone who would immediately slip into the jargon we use among ourselves. I wanted someone who could translate our work into clear, accessible language for a broader audience.

It’s hard to do that when you’ve been immersed in the technical soup for 35 years, as I have. If you’re starting a quantum computing company, chances are you’ve just finished a PhD in physics and you’re ready to build machines. If you’re starting a cryptography company, you’ve probably been steeped in algorithms and key management your whole life. And when that’s your world, it becomes difficult to step outside of it and explain your work in terms your mom or dad would understand.

That’s the challenge we face, and it’s the challenge you face as a journalist. Taking these incredibly complex, technical topics and rendering them into something that makes sense to the people who need to make critical decisions for our collective future. It’s not just important; it’s essential.

Despite the complexity, I see this as one of the most exciting times to be alive. We’re navigating unprecedented technological transitions, and we have the chance to ensure that the world of tomorrow is secure, trusted, and resilient. We really are living in one of the greatest eras.

Although you did not directly ask me this question, let me share what I would do if I could make one critical strategic decision for Canada today.

I would launch a national initiative called “Quantum-Safe Trust Infrastructure.” Like Churchill Port or the Trans Mountain Pipeline, this would be a nation-building project—except this one secures the cyberspace underpinning our digital economy and critical infrastructure.

This initiative would include:

• Establishing a strategic reserve of ready-to-deploy quantum-safe hardware.

• An immediate transition of all federal IT systems to quantum-safe standards.

• A mandate for the Bank of Canada to lead the migration, followed by the broader banking sector.

• A national program to work with provinces and critical infrastructure owners to migrate their IT and OT systems without delay.

• A workforce training and incentive program to ensure Canada has the talent required to deploy and maintain quantum-safe technologies at scale.

The foundation for this project already exists. It is 100% Canadian, and it represents the world’s most advanced capability in this domain. We have the technology, the people, the manufacturing capacity, and the expertise to act—right now.

By treating cyberspace as a new dimension of national defense—alongside land, sea, and air—investments in this project would also count toward Canada’s NATO commitments. More than that, Canada could position itself as the Quantum-Safe Trust Infrastructure provider of choice for all our allies.

This is a “now” project. One that secures Canada’s leadership, sovereignty, and trust in the quantum era.

Key Takeaways

A rare second chance

As I explore the Canadian innovation space, I often find myself surprised, not only at the pioneering contributions that Canadians had made to the technologies that quietly operate in the background of the modern world, but at how unassuming Canadians are when they discuss their contributions. Bruno tells the story of contributing core intellectual property (IP) to the foundations of digital security, IP that has been in productive use for three decades and counting, and he tells it without any pretentiousness.

I was less surprised to learn that Canada no longer controls the IP he developed. The LunaHSM IP has been sold 4 times over its 30 years in service, first to US-based companies, and finally to Thales in France, generating enormous value in the process, value not enjoyed by Canada. In a bit of a break from the usual culprit, though, Bruno’s story doesn’t point to a specific policy failure as the reason for the original sale, simply a combination of timing that coincided with the initial semiconductor hype cycle, relative inexperience as a first-time founder, and investors chasing hype.

Unlike most similar stories, however, Canada has second chance to get it right. Crypto4A is trying to do for the internet what Chrysalis and the LunaHSM did for the nascent internet: provide a hardware level foundation for digital trust, but in a world in which quantum computers are the norm.

Digital sovereignty and nation-building

There is a lot of talk of “digital sovereignty” in response to the realization that American law trumps Canadian law for data controlled by American big tech. PM Carney has promised major investments in a Canadian sovereign cloud. The Canadian sovereign cloud could be a nation-building project or a campaign slogan, depending on whether or not we get the details right. Bruno’s commentary makes it clear that for this to work, Canada must control the keys through which all of this is secured. In his words: “no key, no sovereignty”.

While quantum computing is not yet here, Bruno’s commentary (and 30 years of real world demonstration) make clear that the hardware solutions that are deployed today to secure these keys must be ready for its arrival. Like the LunaHSM, many of the HSMs deployed now will operate for decades and be secure not just against today’s threats, but also those of the foreseeable future. No matter where you fall on your prediction of when quantum computing will be a reality, not even the most pessimistic estimates put it outside the operational lifetime of HSMs that are deployed today. The HSMs that secure the Canadian sovereign cloud must be quantum resistant.

Transitioning to quantum safe standards will have to happen sooner or later, and there is no downside to “sooner”. In Bruno’s words: “Someone is going to do it, because otherwise the internet stops working.”

The benefits are many-fold: getting ahead of a security threat that we can clearly see coming, if not exactly when; procuring solutions from a Canadian company that has the track record to prove their value while directly supporting a key nation-building effort focused on digital sovereignty; and providing opportunities for upskilling of Canadians across every sector of the economy that connects to the internet (i.e. all of it). All of this fits under the umbrella of increased defense spending mandates, and positions Canada as a leader among NATO allies with respect to quantum-resistant cryptography. As Bruno puts it:

“Canada has a chance to turn this into a national advantage. We’ve always punched above our weight in cybersecurity. If we make quantum-safe migration a major national project, Canada could cement itself as a global leader for decades to come.”

Last but not least, it provides a direct path for Canada to secure the economic benefits of world-leading IP in a technology will sooner or later be needed by the entire world. We already lost the LunaHSM, let’s make sure we don’t let that happen a second time.

Geopolitical and technological necessity set the stage for economic and security opportunity. Bruno’s commentary provides the roadmap for Canada to reap the benefits.

Many thanks to Bruno and the Crypto4A team for taking the time to do this interview and share their insights.

In the year and a half I have been ranting writing about innovation policy, I’ve had many opportunities to learn. A comment on one of my older posts prompted me to reread some of my early content, and it is clear that some of it is overdue for an update to take into account how my thinking on these issues has changed.

In the article in question, I came to the conclusion that inventor-owned IP policy in universities is correlated to better outcomes with respect to entrepreneurship than policies in which the institution takes ownership. In hindsight, this was an overly simple take. It came primarily from my own experience with tech transfer and was based on an overly Canada-centric view. At the time I wrote that article I had not yet explored how tech transfer is done abroad, which suggests a more complex interaction between policy and outcome than can be inferred from Canadian examples alone. For example, the United States has almost uniformly adopted institutional ownership of research IP, and has led the world on tech transfer over the past several decades. Clearly, IP policy at the level of universities is not the only thing at play.

David Durand, TJ Misra, and I recently published a related op-ed for IRPP Policy Options that presents a much more nuanced view of the issues at play. Focusing on the Bayh-Dole Act, an inspired piece of American legislation which shaped American tech transfer, this op-ed serves as a spiritual successor to the post linked above. I suggest reading the op-ed (it is not very long) before going any further.

Read the Op-ed

This is the first, but certainly not the last, old article that will get a refresh to take into account what I have learned through the debates and conversations in which this blog has enabled me to take part.

I am going to leave the old article up, warts and all, with a link to this one and a note that it is outdated. I welcome and encourage public debate on these topics, and I think that having a public, written record of the thought process behind these debates is both useful and important.

Context Matters

The op-ed makes a key point that is the core of where my original post went wrong: that we cannot generalize the outcome of a policy framework without also considering the context in which it occurred. While we point to the Bayh-Dole Act as a key example of successful policy on emerging technology commercialization, we also acknowledge that attempts to recreate that success in other contexts has mostly failed: Finland tried, with almost an opposite result, and other attempts have had mixed results, none of which come closes to the American version. New Zealand, inspired by Waterloo, just went in the opposite direction and changed its universities’ IP policies nationally to favour inventor ownership of IP. It will be interesting to see how that plays out, but my prediction is that it makes little practical difference, given that there appears to be little consideration of context in the language around the change.

In other words, the key correction to my original post is that while policy can set target outcomes and provide guardrails with respect to how they are achieved, that alone is not enough. What determines how intention translates to outcome is the context in which that policy is enacted. Canada has a lot to learn from Bayh-Dole, but it has just as much to learn from previous failed attempts to replicate it, and from thinking carefully about the differences that exist between Canada and the United States and how to compensate for those differences in any attempt to recreate its outcomes.

The United States has risk-tolerant private capital that is happy to invest early in the development process, understanding that most of those bets will fail. Canada does not. The United States has a private sector willing to take big swings on new technology, fuelled by decades of proof that technological innovation is the key to economic prosperity and security. Canada does not. The United States has whole-of-government leadership in the form of federal frameworks that support emerging technology (the SBIR, DARPA, Bayh-Dole, and others) which align all the stakeholders toward common goals. Canada does not. Most places where policies similar to Bayh-Dole have been attempted lacked these key ingredients, as well. The lesson here is not complicated.

Kevin Carmichael at the Logic suggested recently that because the Canadian government is making big bets, the private sector should follow suit. He’s right, but private sector investment in Canadian innovation is probably a few years away even if the Carney government delivers on these promises. The private sector will invest when the innovation and competition created by embracing risk and actively supporting emerging technologies forces them to adapt, and that will take time. If Bayh-Dole-like policies are to be adopted here, and I believe they should, we must actively fund and resource the initiative with risk-tolerant public capital, at least long enough for the effects of a more innovative ecosystem to spur Canadian private firms to act in self-preservation, or to fail and make room for those that will.

Industrial policy recommendations from IRPP

The op-ed linked above is a small part of a set of points I made in a round-table discussion hosted by the Institute for Research on Public Policy (IRPP) about a year ago. The round-table was part of a much larger effort to synthesize evidence relating to industrial policy into a set of recommendations for the Canadian government to address Canada’s productivity challenges.

The findings of the discussion series are now summarized in their report entitled Building for the Future: How Industrial Policy Can Strengthen Canada’s Economy and Sovereignty. It is full of gems and is worth a full read through, and I have to commend IRPP on coming up with a report based on a research process that straddled the second coming of Trump while maintaining any kind of relevance.

The report is a solid starting point for debate around industrial policy initiatives that will be necessary for Canada to remain globally competitive, and gives careful thought not just to implementation, but to ongoing evaluation and governance. As with every single remotely related report ever produced on Canadian innovation, it calls for strategic coherence across all levels of government combined with rigorous evaluation and an ability to course-correct as basic requirements of effective industrial policy.

Probably my favourite quote from the report is from Dani Roderick:

“What determines success in industrial policy is not the ability to pick winners, but the capacity to let the losers go”

Substitute “emerging technology” for “industrial policy” and you have the core message of CanInnovate in a nutshell.

What is implied by the quote above, and what I try to get across in some form in most of my writing, is that successful industrial (and innovation) policy requires a high tolerance for risk and the ability to place many bets, accepting that some will fail. In other words, the report makes clear that none of this works without capital, and suggests numerous means by which productive investment can be encouraged.

What’s next?

The lesson about context generalizes: in seeking to learn from any policy framework that has seen success elsewhere, including industrial policy, we must consider the context in which that policy worked and then either adjust the policy to compensate for any differences, or work eliminate those differences. Successful policy import requires beginning with the end in mind and working backwards to identifying the how. The path to similar outcomes may not be the same.

As Canada follows the rest of the world toward a more active industrial policy, we would do well to keep this in mind. We have no shortage of evidence, studies, and good ideas. What remains is action. From our own op-ed:

“We know exactly why we’re failing and we can’t let ourselves off the hook.”

Last year the Trudeau government began a consultation process on reform to the Scientific Research and Experimental Development (SR&ED) tax credit program that incentivizes private sector investment in R&D through use of refundable tax credits. The program is among the largest such incentive in Canada, amounting to some $4B in credits annually over the past few years.

At the time, I responded to the consultation through an initial open letter, and then again in response to the government’s follow up that synthesized what they had taken away from feedback in the first round.

On August 15, the Carney government released draft legislation toward enacting some of the updates, with no substantive changes from the original plan. While the changes were mostly positive, they were mostly low-hanging fruit and did not engage in a substantive way with what seem to be fairly straightforward structural changes to the way the credit is delivered that could make SR&ED more efficient at delivering value per dollar. My view on this has not changed much.

In this article, I reiterate part of what I proposed last time as additional measures that could be considered to maximize the impact of SR&ED and reduce inefficiencies.

What Stayed the Same (Pretty Much Everything)

Everything proposed in December is still there, and as far as I can tell, there is nothing much new. The changes announced in December 2024 indicated that expenditure limits on which the enhanced 35 per cent rate could be earned from $3 million to $4.5 million, as well as increasing from $10 million and $50 million, to $15 million and $75 million, respectively, the taxable capital phase-out thresholds for determining the expenditure limit. These threshold updates are unchanged in the new draft legislature, and represent only an adjustment of the limit to account for inflation since these thresholds were previously set. There is still no measure to continuously update these thresholds to keep pace with inflation. The December announcement also included a provision for extending eligibility for the enhanced 35 per cent refundable tax credit to eligible Canadian public corporations, and allowance for capital expenditures to be claimed for deduction against income and investment tax credit components of the SR&ED program. These are still there, and you can read my thoughts on it in my previous article on the topic.

In other words: the Carney government is moving forward with SR&ED reform as originally planned by the Trudeau government. While I have no objections to any of it, the inefficiencies that exist within SR&ED can only be addressed through changes to how the program is accounted for and delivered. Updating the base numbers is welcome, if overdue, but leaves on the table opportunities to increase the impact of the funding without increasing the cost.

What Could Still be Done

Anyone who has gone through the rigmarole of claiming SR&ED will tell you that the process is deeply inefficient. Claims take months to review, and payment of the credit can be delayed significantly. Because of the complexity of the rules around qualifying expenditures, most companies hire dedicated consultants to prepare their claims, which are usually paid on contingency, taking anywhere from 10-18% of the credit as payment and reducing the capital that flows back to the innovative companies that are being supported.

These structural issues disproportionately impact small companies. While an established company can weather a denied claim, for a small company, especially a pre-revenue company focused on the R&D required to commercialize something truly novel, a denied claim could be an existential threat. The cash-flow problems arising from the timelines and lump-sum nature of SR&ED payments require that innovative companies do one of two things: either they take fewer risks and spend less on R&D in the early stages to ensure that a denied claim will not end them, or they turn to SR&ED lenders to offset cash-flow, paying high interest rates in exchange while increasing the potential damage of a denied claim. Between contingency-based SR&ED preparation fees and interest rates on SR&ED lending, these cottage industries siphon off a large percentage of the credit. If this could be addressed, and if SR&ED could be better aligned with cash-flow requirements, the impact of this credit as an enabler of intelligent risk-taking would be dramatically increased without increasing the cost to taxpayers at all.

The means by which this could be achieved are also not complicated in principle.

SR&ED claims work in most cases through what is called the “proxy method”: instead of accounting for and claiming specific R&D expenses, companies can simply estimate the total salary paid in service of R&D, add 55% overhead, and claim that total as eligible expenditures (the calculation is a bit different for contractor costs, but we will focus on salary). The obvious change to SR&ED delivery is baked into the proxy method: because SR&ED credit is directly proportional to R&D-related payroll, SR&ED should be integrated with the payroll tax system.

Payroll tax is paid monthly. By incorporating a simple calculation of the percentage of each payroll tax submission that is SR&ED-eligible, the SR&ED credit could be used to first reduce the amount of payroll tax paid each month and then credited on a rolling monthly basis. Companies could then be required to submit the usual annual report that provides an overview of the past year of expenses as a means to verify these payroll tax deductions, but because this would be a correction to payments already made, the cash-flow implications would, assuming no major errors, be minimal.

This change alone would completely eliminate the SR&ED lending industry.

The second structural change should be focused on reducing uncertainty and the chance of a denied claim. Instead of post-vetting of research expenditures, pre-vet them. The annual report mentioned above should include not just verification of the payroll tax deductions claimed in the past year, but also a research plan for the following one. CRA could then provide advance notice to companies of ineligible expenses, which would allow for proactive adjustment to correct for any issues before they arise. The work required by CRA remains effectively the same, since they are approving these research plans regardless, but moves from a reactive, post-approval model that risks unexpected denials to a proactive, pre-approval model that reduces instances of denied claims. Increased confidence in expense eligibility and predictable cash-flow in turn promotes intelligent risk-taking and R&D-based innovation by Canadian firms, eliminates entirely the need for SR&ED lending, and reduces the value-add of SR&ED consultants, ensuring that more of the tax credit is used for its intended purpose. Together, these are potentially cost-neutral reforms that increase the fraction of the SR&ED credit that ends up where it belongs: in the hands of innovative Canadian companies.

Next Steps

While an exact estimate is somewhat challenging given public data on the topic, it is reasonable to estimate (as discussed in my previous posts on the topic) that upwards of 20% of SR&ED spending is not ending up where it should: in the hands of innovative Canadian companies,. Parts of the credit are being diverted through SR&ED consultants that exist only to navigate the complexity of the related administration and lenders that exist only because SR&ED is paid out as an annual lump sum rather than a monthly payroll tax credit. Addressing these through relatively simple changes to the way the credit is paid out could greatly increase the overall impact without changing the cost of the program.

I stand by my original recommendations as well, especially those relating to ensuring the SR&ED does not go to supporting subsidiaries of foreign firms. I will not reproduce them in full here, simply suggesting that you take a moment to review my previous article on the topic.

I intend to submit suggestions along these lines in response to the feedback invitation in the hope that the current administration is a little more ambitious than the last.

Several months ago, the office of Senator Colin Deacon produced an excellent report reviewing federal innovation funding initiatives, finding nearly 140 programs that aim to support innovation in Canada. As someone who makes a point of staying on top of Canadian innovation trends, I had heard of less than half of them.

While I am happy to see additional resources being earmarked for innovation via defence spending, the search for efficiency through deep cuts to the size of public sector before first finding it in underperforming innovation programs suggests that we have somehow still not learned the critical lesson.

Recently, the Information Technology & Innovation Foundation (ITIF) picked up the thread in the form of an excellent op-ed by Lawrence Zhang. Today I draw your attention back to this critical debate, highlighting his analysis and re-examining a well-established issue in light of the priorities of a new government.

Policy fragmentation

Lawrence Zhang’s assessment of the federal Canadian innovation ecosystem is blunt, and more or less echoes themes that thread through most of the articles in my own work:

“For decades, we’ve treated innovation as a policy file, not a system. We’ve staffed it with generalists and managed it through siloed programs and compliance. The result isn’t progress, it’s paperwork. No agency is mandated to steer strategy across departments, sequence support, or commercialize. […] These programs matter, but they operate in isolation—stranding firms, fragmenting funding, and stalling technological advancements.”

This is a dense paragraph that points to many problems. “Staffed with generalists” references a perennial challenge with the public service arising from a mismatch between the experience of many program administrators and the realities of the programs they administrate (a topic for another day). Endless paperwork bogs down the system, causing innovation support programs and to move too slowly to effectively support innovation. A lack of cross-cutting mandate and “whole-of-government leadership” leads to isolation and fragmented funding

This fragmentation has many consequences. As Lawrence points out, it often leaves firms, especially early-stage and pre-revenue firms, stranded in the gaps at precisely the stage where support is most needed. Even when funding is available, the complexity of access stretches timelines beyond the timescale required for innovation Canadian technologies to be globally relevant. I recently wrote about the opportunity costs of these issues: a prime opportunity for Canada to reverse its brain drain and benefit from America’s misguided recent research policies, likely to be completely missed because we are not prepared to capitalize on their mistake fast enough to capture any of the value.

I would add one more core issue to the list: an unwillingness to cut existing programs.

The result is what we have today: 134 fragmented programs. While there are no doubt gems among them (I have written previously about the good that Lab2Market, Mitacs, and organizations like IPON are doing for emerging technologies, for example, and while there is much that could be improved about SR&ED, I suspect it does more good than harm), many of the existing programs do not collect or report data that correlates specific programs to long-term outcomes, and an overreliance on job and revenue-related metrics prevents an objective assessment of program efficacy.

As Lawrence points out, numerous proposal have been tabled to address the issue, with the current thinking around the issue being established in 2011 with the Jenkins Report and repeated in the Bouchard Report. In an echo of the innovation policy frameworks it sought to change, the report generated mostly paperwork and little progress, though not for lack of excellent recommendations. Over the subsequent years many solutions have been proposed: the CIC was floated as an agency intended to streamline and coordinate scientific funding across the three main siloed research funders (NSERC, CIHR, and SSHRC), but was delayed and is now unlikely to see the light of day, even though I still hear about it occasionally on conversation with policy makers. The ambitious CARPA (modeled after DARPA and proposed in the 2021 Liberal platform) met a similar fate. The proposed capstone agency for tri-council funding appears to be in bureaucratic purgatory.

Shortly before the ITIF op-ed was published, the Carney government announced the Bureau of Research Engineering and Advanced Leadership in Innovation and Science (BOREALIS), which appears to be the latest in a long line of ambitious announcements. I am reserving judgment on the follow up for the time being.

Lawrence proposes a combination of past attempts:

“What’s needed is a combination of the two proposals—an institution that takes responsibility for the system itself. That’s why we need a Canadian Innovation and Industrial Transformation Agency with the authority to coordinate strategy, steer delivery, and close the structural gaps that hold back deployment.”

Lawrence does not directly address BOREALIS, but much of what he advises can be directly applied to any attempt to execute, especially his admonition that any such program aimed at delivering long-term innovation must be at arm’s length from the government:

“It must be a Crown corporation. Not a secretariat or special office. Only a Crown can operate outside standard government HR, IT, and procurement constraints while maintaining strategic accountability to Ministers.”

He suggests that the new agency should coordinate delivery of (at least)

“National Research Council’s Industrial Research Assistance Program (IRAP) and its 350 staff; the Strategic Innovation Fund (SIF); Innovation Solutions Canada (ISC); the Department of National Defence’s Innovation for Defence Excellence and Security (IDEaS) program; and design and appeals functions related to the Scientific Research and Experimental Development (SR&ED) tax credit”.

This would be an excellent start, and would fit the definition of “do more with less” that has been the hallmark of what PM Carney has asked, directly or indirectly, of the various programs under funding review.

Demolish first, build after

While I agree with pretty much all of Lawrence’s advice on what Canada needs to change about its approach to innovation, There is perhaps only one point on which I disagree. Lawrence asserts that “This isn’t about replacing programs. It’s about governing them coherently, strategically, and at speed. ”

Whatever is built next, be it the CIC, CARPA, BOREALIS, or Lawrence’s Canadian Innovation and Industrial Transformation Agency, I think it should be about replacing programs. 134 federal innovation programs are not something to preserve for the sake of avoiding stepping on toes, and very little in the current system is able to “govern coherently, strategically, and at speed” in the current siloed approach. It is abundantly clear, and has been for at least a decade, that what we are currently doing does not work.

Answering the question of what to cut, what to reform, and what to merge is of course much harder than the suggestion to do it.

We should not treat any program as sacred. My litmus test for what should be cut or repurposed revolves around the quality of the data on which an objective assessment can be based, and the degree to which it is explicitly and intentionally connected to other, complementary programs.

A program for which value cannot be demonstrated quantitatively (or at least concrete progress toward value creation, given timeline considerations) from metrics that same program collects should probably be slated for at least reform of how it collects and reports impact. A program that does not obviously connect to both upstream and downstream support should probably be restructured to explicitly align with what comes before and after to ensure no gaps. A narrowly scoped program that provides funding for just one piece of the puzzle without working hand-in-glove with complementary programs should have their mandate updated to be an active participant in ensuring that support recipients have access to the rest of the resources they need. A program that is unwilling to engage with pre-revenue companies should have its risk tolerance recalibrated.

We are being asked to balance conflicting goals that will play out on different timescales: build a self-sufficient Canada powered by innovation, invest in defence and the NATO Alliance, and at the same time tighten our belts to weather the short term economic storm caused by our neighbors to the south. As defence spending ramps up and Canada seeks to build economic resilience and reduce economic dependencies, we must take a long view of what it will take to achieve this. Innovation today lays the foundation for economic resilience and security tomorrow.

If we work to increase cohesion and demonstrable impact across the board, efficiency will follow.

While at QUANTUM NOW | ICI QUANTIQUE, I had the opportunity to interview senior representatives of three leading quantum computing companies, including Rafal Janik of Xanadu, Paul Terry of Photonic Inc, and Allison Schwartz of D-Wave about the future of Canada’s quantum computing ecosystem, and the implications of DARPA’s Quantum Benchmarking Initiative (QBI) program.

My goals with this article are three-fold. First, I aim to set the scene and highlight the urgency with which Canada must take action if we are to remain relevant in quantum computing. Second, I try to turn these conversations into actionable recommendations for policy initiatives intended to ensure that Canadian quantum computing companies are able to stay Canadian as they cross the finish line. Finally, I hope to make clear the value of actually following through.

While Photonic and Xanadu are both currently Canadian quantum computing companies, D-Wave is no longer. Founded in 1999 in Vancouver, D-Wave operated as a Canadian company for 23 years before going public on the NYSE in 2022, while maintaining an active presence in Canada. In my view, their move represents a failure by Canadian policy makers to act sufficiently quickly and signals the beginnings of a trend that must be proactively halted if Canada is to realize the benefit of its decades-long investment in quantum.

Since this article draws on multiple interviews, I am going to switch up the format a bit and present the key takeaways first, relying more heavily than usual on pull quotes. The full interview transcripts for Photonic and D-Wave are presented in separate sections below, while Xanadu requested that I limit publication only to pull quotes. As with all my interview-based posts, it is too long for email and will likely be truncated by your email client. For the full effect, you will need to read it online.

Many thanks to all of the interviewees for their generosity with their time and for sharing their insights.

Interviewer’s note: All interviewees reviewed and edited the the interviews and quotes posted below, with the option to rephrase and expand on the ideas discussed without changing or removing any intended meaning. The key takeaways are my own commentary, were not reviewed or edited by interviewees, and do not necessarily represent the views of any of the interviewees or their respective companies.

Key Takeaways

The DARPA Quantum Benchmarking Initiative

DARPA QBI intends to deliver a “utility-scale” quantum computer in 8 years, or prove that it can’t be done. This particular DARPA initiative is odd in that it is focused on delivery of a commercial technology for civilian use, in contrast to previous DARPA programs that were defense-focused and in some cases actively discouraged or forbade commercial translation. This is consistent with recent trends toward a reversal in the order of operations with respect to emerging technology development, with a recent NATO report highlighting that, in contrast with the historical approach that has characterized military technology development since World War II, emerging technology development is increasingly occurring in the commercial private sector first.

DARPA is focused on gate-model systems, which excluded D-Wave from consideration, as its commercial offering currently is a quantum annealer. As Allison describes:

“[…] quantum computing is not monolithic. There are different modalities or different qubit architectures. Each of these are advancing on a different timeline, and each of these have different strengths and weaknesses and capabilities. Annealers are best for optimization problems, which are ubiquitous across industry. That's what's commercially available [via D-Wave] to now […]”

Among the 18 companies globally that were selected to participate, four are Canadian (Xanadu, Photonic, Nord Quantique, and 1QBit). The program is divided into three phases (A, B, and C). Stage A is a 6-month process in which the companies present a large amount of information relating to the details underlying their delivery plan and the technology on which it is based. Stage B lasts 12 months and involves technical due diligence, with the DARPA team going through the underlying research and the proposed execution plan with a fine-toothed comb. Stage C involves execution of the subset of plans deemed worth funding, and, eventually, actual delivery of a utility-scale quantum computer.

The DARPA Dilemma

Regardless of the success of the program and acceptance into Stage C, participation is a risk for the companies themselves. Paul explains the dilemma:

“For the people that don't go through, is that a statement that their technology doesn't work? Is that a statement that, politically, they are people that are not meeting the US requirements? Those requirements could be wide-ranging. So for example, those requirements could have IP constraints and restrictions. They could have US domestic considerations. They could require you to move to the US. […]

It could be that the technology is not why they do or do not get into DARPA. If you speak to Rafal or Julian, they’ll tell you the exact same thing: that we worry about the signal of getting in, or not, to DARPA-C. Politically, you could take the view that people going to DARPA are pandering to the American stuff, or you could take the view that the technology wasn't worthy. It's a “Sophie's choice” from our perspective, we can't actually make the right decision.”

In the end, the decision comes down to the value of third-party validation and the opportunity to learn in turn from DARPA, and use what Paul refers to as a “forcing function” on the technology, using the questions posed by the DARPA technical team to learn and advance in turn.

What came through clearly, both in panel discussions at the conference and in the interviews, is the depth of due diligence being conducted by DARPA, with weekly half-day meetings between dedicated DARPA technical teams and senior company leadership, hundreds of DARPA analysts, and reams of technical documents changing hands. The program's depth of inquiry suggests that the Americans are benefitting already. As Paul puts it:

“They're learning a lot, actually. I can say from personal experience, things that they took to be true, they're learning are not necessarily true. It’s actually accelerating, not decelerating, the perception that quantum is going to be here sooner rather than later.”

While the information changing hands in stages A and B does not constitute details of the IP required to make current systems work and related IP terms were uniformly viewed as low risk, to my mind the main risk to Canada lies in DARPA Stage C. Performance requirements under Stage C have yet to be defined, but Paul’s comment above indicated that requirements to be physically located in the US are not out of the question. This is supported by indications that testing of the results delivered under the QBI will be conducted at least in part in facilities in construction in Maryland, and the investment of up to $300 million per successful company through DARPA—on par with Canada’s entire National Quantum Strategy Budget—combined with the possibility of anchor customers and third-party validation, will exert a strong pull.

Building a Quantum Sandbox

In 2022, Allison addressed Canadian Parliament with a proposal to build a “Quantum sandbox,” a set of infrastructure, hardware, and tools on which to base quantum and hybrid computing education, application development, and value demonstration. As Allison describes it:

“I testified before Parliament almost four years ago. I pitched the idea of a quantum sandbox and everybody thought it was a great idea. It was actually a recommendation from Parliament to the administration and we're still waiting for it to be implemented.”

The testimony became a recommendation in the resulting report, and subsequently went nowhere. In the meantime, the Americans are on it.

This is unfortunate for many obvious reasons, and a few less obvious ones. One of the highlights of the interview set was a comment by both Allison and Paul about the importance of the application layer and the value of everything that is enabled by quantum computing hardware, as opposed to the hardware in and of itself:

Allison:

“Where Canada can grow is on the application layer of seeing where these actually can do useful things. We need both. We need both the investments in the hardware advancements as well as the software, and we need investments in all the different modalities that are out there.”

Paul:

“We have an opportunity here to create what happened in Silicon Valley 30 years ago in Canada. I'm actually more excited about the surround, the quantum companies in the application space, around the quantum computers that we have here, more than that than the quantum computers themselves. Quantum computers themselves are, to a large extent, a solved problem..”

Their observation is critical: quantum computing could catalyze Canada’s Silicon Valley moment. If you have read any of my previous articles on ecosystem building, or Dan Breznitz’s book on innovation, you will find a lot of discussion of the magnetic pull of established innovation ecosystems that draw value away from smaller ones. The existence and gravity created by Silicon Valley is one of many reasons it is difficult to build something similar in Canada. Where innovation ecosystems are concerned, first mover advantage is everything, with a caveat: to some degree, the effect is sector-specific. With quantum technologies, Silicon Valley has limited gravity. It has no critical mass of quantum computing talent nor an established base of research, which represents Canada’s best, if not only, chance at creating that gravity here. We may not be able to compete on a pure cash basis, but we can use our current critical mass of expertise to balance some of the difference.

The value creation of Silicon Valley is much less about the anchor companies themselves as it is what was built around them. The software boom that has been the core of VC for the last few decades was built up on infrastructure. A quantum sandbox, built on Canadian quantum computing hardware, would have wide ranging knock-on effects, starting with the infrastructure required for a generation of hackers to develop quantum algorithms and create value by enabling practical applications.

For those of us still hung up on job numbers, Paul makes clear that this is where the vast majority of job creation potential lies, as well:

“I often get politicians saying that, ‘in Photonic you're creating 300 jobs. 300 jobs isn't interesting.’ Well, it's interesting to me because these are super PhDs. These are the best in the world, people. They create five jobs for every job I put in here, but more importantly, the application surround around this is going to create thousands of jobs and change the GDP of a nation, which is not necessarily a function of the number of jobs that you have.”

Execution will require large, public procurement projects aimed at Canadian quantum computing hardware companies, thereby providing them with anchor customers and signaling robust commitment from the Canadian public sector. Procurement need not be limited to the actual hardware. The federal government could also commission a Canadian quantum company to develop a dedicated quantum-optimized logistics solution for a major crown corporation, secure quantum network for a defense department facility, or optimize logistics for allied military operations. These projects are already well within the capabilities of present-day, commercially available quantum computers.

No matter the details, it has to be led by the public sector, a theme that comes up in many of my articles. As elaborated later in this post, this is not a request for grants, but for anchor customers, customers that are still overwhelmingly in the public sector globally. As Rafal points out, quantum computing is heavily subsidized everywhere:

"You'll see that this is not a free market. This is very heavily subsidized, everyone has a thumb on the scale. Historically Canada has said it's not going to pick winners, ‘there's going to be a free market on this, so we're not going to do it’ It just means that Canadian companies are fundamentally at a disadvantage."

Given the dual use nature of quantum computing, all or almost all of these expenditures could fall under the recently announced increased defense spending under BOREALIS, Canada's strategy for critical minerals and Arctic sovereignty, which directly relates to advanced technology infrastructure, including quantum.

Security Considerations

There are economic security arguments for doing this as well.

While there are real IP risks associated with the DARPA QBI program, they are not the only risks. Globalized supply chains, in particular where chips are concerned, represent key points of vulnerability. As long as critical components are being manufactured abroad, there is a risk of IP theft.

Lisa Lambert of Quantum Industry Canada recently wrote an op-ed in the Globe and Mail calling for industrialization of quantum technologies in Canada and “fabrication, integration, secure supply chains and deployment at scale.” Having Canada as the NATO epicenter of quantum technologies allows us to begin the onerous and expensive but increasingly necessary process of deglobalizing supply chains while making ourselves invaluable to the alliance.

Paul also highlighted a key issue relating to export controls. Quantum technologies are an export control nightmare, and the country that hosts related services will be in a powerful position to dictate how regulatory development unfolds:

“The growth of Quantum and the understanding of what Quantum is going to do to nations has created quite an export control problem for everybody. […] That's if you're selling hardware, lasers, semiconductors and stuff between countries. What if you're not selling that and you're selling Quantum services? […] I imagine that the Canadians or governments will have something to say about releasing Quantum services of this power to anybody in the world,”

We have been here before: in 2017, Canada was a world thought leader in AI, and the Montreal Accords were recognized at the time as being a valuable contribution to the regulatory debate of global value. Today, I doubt many people outside of Canada even remembers them. With quantum technologies, and quantum computing in particular, it is not too late to replicate for this sector what we tried and failed to achieve with AI, and from across a much deeper moat. Establishing Canada as the centre of quantum computing hardware, application development, and the manufacturing supply chain needed to make utility-scale quantum computing a reality will also allow us to lead the global debate on regulation with respect to export controls, quantum service provision regulation, and standards development, all of which are critical elements of economic security.

Now or Never

Time is limited. Paul points out that a consolidation of the market is coming:

“There will be a consolidation at some point in the next four years, I would think, of Quantum companies in the world.[…] There's two hundred and something Quantum companies in the world. In three years there may be four.”

This is supported by a recent raise by D-Wave of $400M in preparation for acquisitions. Between impending consolidation and the 18-month timeline for DARPA Phase C to kick off, Canada has about 12 months to get its act together. To do this, the federal and provincial governments need to buy quantum computing hardware from Canadian quantum computing companies, and keep buying more as it improves. Use it to build the sandbox we should have built five years ago, and enable Canadian startups, researchers, and hackers to start building the application layer on hardware that is ready to use today, building the application while at the same time training new talent, providing a reason for existing talent to stay, and for talent abroad to immigrate. Establish Canada as the centre of quantum gravity for NATO, and use all of that as impetus to industrialize as Lisa Lambert suggests. Be the anchor customer that keeps the ecosystem here.

Either we act before DARPA Phase C commitments are made, or we can assume that quantum will go the way of Canadian AI, yet another Canadian technological gift to the world, still debated long after we lost the thread. Rafal and Paul have effectively the same message to share:

Rafal:

“Canada needs to decide, ‘do we want sovereign quantum computer capacity’? If the answer is yes, it needs to get very serious about what level of investment is required to actually deliver that.”

Paul:

“The irony of all of this is there are only four thousand people in the world or so who are Quantum experts. So those four thousand people are going to appear or reappear in the one, two, or three Quantum companies that are going to win this. Canada has a chance of having one, two, or three of those companies based in Canada, from a hardware perspective.”

[Interviewer’s note: Paul’s 4000 number apparently refers to people in “Canada's commercial and academic quantum talent pool” from this report, which as far as I can tell is an accounting of Canadians with PhDs in quantum physics. The global quantum workforce beyond PhDs is far larger, as he confirms in the interview.]

If we want to capitalize on 25 years of investment in quantum and a world-leading position on a technology that has been established in Canada, we need to get it done before DARPA Stage C.

As to what comes next, it is clear from all three interviews that the ball is firmly in the court of Canadian policymakers. As Paul puts it:

“Those companies, by the way, are fiercely Canadian. So this view that we are in it to sell to the highest bidder couldn't be further from the truth. We are companies, we do have cap tables, we do have investors, we are subject to capital flows. But speaking for Photonic, Photonic is two thirds owned by Canadians, and it's fiercely Canadian. We’re banging the table going, ‘I want to stay Canadian. Give me a chance to stay Canadian.’”

The will and the money are there, what remains is action. It is up to Canada to make a choice, and it is a very simple choice: Now, or never.

Interview with Paul Terry of Photonic

KB: Let's start with a bit of context. Give me a short intro to Photonic, what you're attempting to build, and what it's taken to get where you are today.

PT: Photonic was founded in 2016 by Stef Simmons, who's a physicist at SFU, a professor there. She had the idea of using colour centres that are photonically interconnected. I I mentored her in 2016-2019 as we calculated the technology necessary to actually sort of win this. So it took three or four years to do the research and to find the attributes of the foundational technology that's necessary to actually make a computer. So in 2019, we were looking for a CEO. Hard to find a CEO that has the experience of a startup physics degree and engineering degree etc. So I stepped in to be the CEO. We actually raised our first round in 2021.

That's where we are now. We're 150 people based in Vancouver, BC, in Coquitlam and we are building the company in Canada. About 30% of the company doesn’t work in the head office and they work wherever they are - across Canada and around the world. That's allowed us to hire the greatest and the best in the world just wherever they live and of course, if anybody wants to come to Canada, when immigration allows it, we bring them into Coquitlam.

KB: Canada has a leadership position in quantum tech currently. What is it about the Canadian ecosystem that made it possible for Canada to be a place where you can build this kind of company?

PT: Canada has long had a culture of funding research. It does not have a culture of funding commercialisation. I've been here for 40 years. This is my seventh company in Canada. I've had five acquisitions and one public company in the last 40 years, and I've been on almost every Prime Minister's committee over the years, on various technologies. I've also had the experience of spending 20 years on the Board of Michael Smith Foundation, plus Genome BC, plus Life Sciences in BC, plus a hospital in Vancouver.

So I spent a long time on the research side and a long time on the commercialisation side. Canada Funds Research Well. Researchers will say it's not enough. and that’s probably true, but then all researchers around the world will say it's not enough and that's probably true too.

But 20 years ago Canada funded quantum. Stef Simmons, the founder of this company, was one of the first graduates of a quantum programme at the University of Waterloo. So she is a direct product of Canada's investment in quantum.

Quantum is a complex subject. To actually work in it, you need a PhD. I think we have 70 PhDs in our company right now. And Canada has them. The UK has them. The US does not have them. So there are various peoples around the world, various countries that have bits of them, because they did not spend 20 years spending billions of dollars on the research. So Canada actually should be credited for doing it, because now you planted the seeds 20 years ago and now the trees are growing. We actually have the best quantum people in the world in Canada. There is really no argument about that at all. So we benefit from that, as evidenced by the fact that we could prove it.

KB: Careful not to let Lisa Lambert hear you say you need a PhD to work in quantum.

PT: Let me just qualify that. You need a PhD to do quantum science and engineering. Having said that, a quantum company is actually quite wide. You need some very strong engineers, you need some good marketing people. In our company we have people with degrees in English, all the way to PhDs in computing. We often have something like a thousand applications for every position in the company, give or take. It's a very competitive place to get into. You don't need a PhD to get in. It depends what job you're going to do.

KB: Photonic is one of four Canadian companies that are part of this DARPA QBI challenge, intending to deliver an enterprise-scale quantum computer by 2033. Tell me a little bit about the challenge that's been put before you and Photonic's part in it.

PT: Let's go back to DARPA. Previously to this I was a CTO of Cray Supercomputing. I've done many DARPA programs in the past. This is not a normal DARPA program. It's the first thing to know. DARPA normally is a defense type thing. They do things to aid the defense of the realm, as it were. This isn't that, which is odd, because DARPA, in this particular case, is solely focused on the commercialization of quantum. Not necessarily the defense aspects of quantum.

You may know that DARPA, of course, funded all the IP networking stuff back in the day. They did a lot of compiler design, a lot of supercomputing design, a lot of things going into space and so forth. These tend to be closed contracts, defense-oriented, with terms and conditions that almost prevent you from commercializing. This program is not like that. This program says, unless you can commercialize, we are not interested in talking to you. That's a very important and subtle difference, and a very big one.

DARPA had a trial run of this program two years ago. Two years ago it said, “before we announce this program, let's work out what the program is and do the trial run with Microsoft PsiQuantum and Atom Computing”. They limited the trial run to only American companies and they took them through the process as a pro-forma for the process they're in now.

DARPA is in three stages, A, B and C. Stage A was a million dollars. Stage B is up to $15 million. And Stage C, there's a promise, not a guarantee, subject to Congress, the Senate and many other political machinations of up to $300 million per company in matching funding for the production of, let's say, a prototype or something that looks like a prototype.

Two of those three original companies are in DARPA Stage C. They have now attracted, 18 additional companies into DARPA Stage A, which is almost a due diligence program before you get to DARPA Stage B.

Notice that DARPA is not funding an end-user computer. It's really about putting a stamp on Quantum, about creating a pivot point for Quantum. Quantum is happening now, so let's fund these companies and advise these companies.

If you're a company in DARPA that does this very well, but doesn't do that very well, DARPA is also interested in doing whatever it takes to make what they call a utility-lifter-scale computer. So as an example, if you had a supply chain issue, if you needed to build a fabrication facility or whatever, they could act as a matchmaker to make that happen. Their sole goal is to make the environment to make a utility-scale computer a reality in a world where people can't easily diligence quantum computing. So it's very much a reality-driven commercial arrow.

To do that, there's a core DARPA group based in Washington, DC, and then there's a Surround group, which is what they call their test and evaluation team. For that team, I think they've hired several hundred people across all the institutions, across institutions like National Labs and so forth, to be the testing evaluation team. Those testing evaluation teams are only for particular companies. T&E teams are not shared across companies. The DARPA main team can go across companies, but they are very, very, very sensitive to confidentiality. A lot of things that we share with them, things they share with us, they keep confidential, even between T&E teams.

They're interested in finding out the integral details, but not sharing it, and in fact, when we send them documents and so forth, those are all marked up with what's confidential, what's not, what they can share, and what they can't. The T&E teams meet with us anyway every week. If you've ever done a PhD and you have an external examiner going through every sentence, that's kind of what it's like. They are very detail-oriented, very keen to learn, and very keen to understand.

One of the things that is a misnomer, is that people think the DARPA team are quantum experts and they know all the answers and they don't. In my experience they are learning, they're absorbing, they're very smart people, very fast, and so there's an educational part to this, to explain to them why this is, why this matters, how this matters, what the ups and downs are. What they're not expecting is for teams to have everything dotted and crossed, what they are expecting is to have a plan, a strategy, a view, a risk profile, and a plan of action to obviate that particular issue.

As an example, if you had a supply chain issue for a bent piece of metal that you need for your quantum computer, then they would say, “okay, so that's good. I recognize you don't have it today. What's your plan to have it, when will you have it, how important is it, what's the second source of it, what's the third source of it?” From a diligence perspective, it's probably the highest level of diligence I've ever seen in my 40 years of doing technology in the world. So welcome to my world.

We have teams of people, normally senior people, who are anticipating the meetings, providing information before the meetings, so that people going into the meeting have the information ready to go, so it's not a presentation every week. It's not a marketing thing, it's a deep technical thing. The best way I could describe it, if you’ve done a PhD, it's like a PhD supervisor going through and trying to uncover all of the details, and they're doing that with all of the 20 companies in DARPA, with different T&E teams.

They're learning a lot, actually. I can say from personal experience, things that they took to be true, they're learning are not necessarily true. It’s actually accelerating, not decelerating, the perception that quantum is going to be here sooner rather than later. Which actually is very good, because no number of marketing materials can beat showing you how QRE actually works, what resourcing means, what a qubit actually does, and then understanding that at a deep technical level. They understand it, even if you haven't done it yet, they understand it, and that actually sets a path for the full thing. The third-party aspect is especially important, because the marketing materials are coming from the quantum companies, and it's powerful having third-party validation.

The other side of that, of course, is that DARPA is a US government object, and you're subject to US government oversight, and nobody knows what they're going to do next, in a whole range of things, to state the obvious, and whether or not DARPA-C is actually real. Someone will say it is, because there are two companies, PsiQuantum and Microsoft, already in DARPA-C, so there are already participants in DARPA-C. So there's an argument to say DARPA-C is already in place, but how many companies will make it to DARPA-C is a question.

Governments are notorious at stating a budget and then cutting it. So will that actually limit the number of people going through? For the people that don't go through, is that a statement that their technology doesn't work? Is that a statement that, politically, they are people that are not meeting the US requirements? Those requirements could be wide-ranging. So for example, those requirements could have IP constraints and restrictions. They could have US domestic considerations. They could require you to move to the US. So this is basically, this is also a threat for all the DARPA-C companies, because once you're in it, you'll do your best, and believe all the Canadian companies will do very well technically, being among the strongest in the world technically. So two questions then to the Canadians: What do you do with Canadian companies that get into DARPA-C? And question two, what do you do with Canadian companies that don't get into DARPA-C?

It could be that the technology is not why they do or do not get into DARPA. If you speak to Rafal or Julian, they’ll tell you the exact same thing: that we worry about the signal of getting in, or not, to DARPA-C. Politically, you could take the view that people going to DARPA are pandering to the American stuff, or you could take the view that the technology wasn't worthy. It's a “Sophie's choice” from our perspective, we can't actually make the right decision. What we do is we make the decision that best showcases the technology that we have and take advantage of the fact that this actually is a two-way street. DARPA also talks to us and goes, “well, we'd like to see this better, or that better, et cetera.”

So it's a forcing function on the technology. Even with the best way in the world, you can't see what you can't see, and so having a third party independent say “you should do that, or think about this, or have a good statement for that” is a very healthy situation. So at the end of the day, the decision that I made, that we made, is that, it's better to be in it and have that forcing function than not be in it. The Canadians need to understand what the future looks like here. And frankly, the British and the Europeans as well, because this has had quite the geopolitical effect on all countries as the US is moving to close. And the US is very good at that, moving to close. It's not, “let's just keep this going and then make a decision in three years.” This thing [DARPA] will all be over in 18 months.

KB: As I understand it, the first two stages of this DARPA program will not be heavily IP-generating and that most of the breakthroughs, should they happen, will be in Stage C. Does that align with your understanding? Has there been a lot of IP transfer in the necessary part of this diligence process? What is the timeline that you expect before the IP-generating breakthroughs need to be delivered assuming that stage C continues and Photonic is part of it?

PT: The IP considerations for phase one and possibly phase two, I haven’t seen the phase two ones yet, are actually very friendly. So they're not at all problematic. We have had experiences in the past with the a government where the IP terms were so outrageous that we just turned them down despite a very significant revenue opportunity. So I agree that stage one and two IP terms look good. DARPA is very keen, and they were very receptive, by the way, to changing IP terms, which we did. On behalf of everybody, nobody signs the same IP term to my knowledge. I don't know if anybody has different IP terms.

The decision to go to DARPA C is a two-way decision. They can offer it and we can decline it. And so if there is only one, two, or three companies that go to stage C, then that will be a negotiated IP thing compared to stage one and two, which was a more general IP thing.

Now, the breakthroughs happened before DARPA started. So the IP terms for matching funding are going to be interesting. I don't really anticipate, and knowing that companies have gone to stage three, that their IP terms have been necessarily problematic. But those are American companies, not Canadian companies.

So we have choices. We do have choices. Stage one and two, good. Stage three, we have choices. And whilst it's up to 300 million per company, it's not 300 million per company you can choose. It's matching funding, so you have to have enough investment in the company to match. Why would a government give you monies without IP terms?

Now, there is a conversation that says that DARPA C wouldn't actually be funding R&D. It would be funding pre-sales, and I don't think DARPA knows yet, frankly, what form it will take. I don't think they have a good view on it. They recognise that they can do it in different forms. They also recognise that it's going to get their money from Congress. So there's a political battle to be had. I don't actually know that anybody knows what the answer is.

KB: Has the current geopolitical chaos substantively changed Photonic’s IP strategy in general?

PT: The growth of Quantum and the understanding of what Quantum is going to do to nations has created quite an export control problem for everybody. There are several agreements. The Wassenaar agreement, there's NATO, AUKUS, Five Eyes, national agreements, trade agreements, all of which have a bearing on what you can do and when you can do it. Can you move a Quantum computer from Canada to England? There are at least two or three agreements that will dictate what you have to do and how you have to do it. So, the government of Canada really has to kind of figure out what it’s going to do with that.

That's if you're selling hardware, lasers, semiconductors and stuff between countries. What if you're not selling that and you're selling Quantum services? You host a computer in a country, could be Canada, and then people access that via the Internet, which is probably the form it will take. The regulations for that today are only privacy type regulations. So, we have GDPR and stuff like that as a regulatory body. That one is going to be interesting. What are the regulations about you using a database piece of software on Amazon Cloud that's hosted in Germany?

I imagine that the Canadians or governments will have something to say about releasing Quantum services of this power to anybody in the world, because people could use those Quantum services to decrypt things, to break passwords, to do stuff. That regulation has not started yet because we don't have Quantum services yet. So there's a problem with hardware today. It's a mess, but that problem will get dwarfed by the provision of Quantum services and whether you have to host a piece of hardware in a country to offer services in that country. That kind of happens today in Microsoft or Amazon, which will actually have different data centres in different countries specifically to stay on the side of the regulatory bodies in those countries. That may happen. So, that's a brave new world. Welcome to my world. That's going to be an interesting one.

KB: It sounds like there is an 18 month timeline for most of the key elements to resolve with respect to the DARPA QBI. What policies or changes does the Canadian government need to make within that timeframe in order to capitalise on the lead that Canada has right now?

PT: The first one is recognition, of course. I would say that having kind of banged on the door of the Canadian government for the last four years, nothing has happened until this change of government and then as we saw this morning they have a Minister for AI and Quantum now. So, this is good. The recognition is good. This event is good. Having the G7 state that Quantum is here and real is good. We met the Prime Minister of the UK on the weekend. The leaders and the political motivations are good. I think we are having the leaders recognise what this is. There are only now three or four companies in Canada that are world leaders in Quantum. I recognise my colleagues in other companies here. There will be a consolidation at some point in the next four years, I would think, of Quantum companies in the world. Recognising that lots of Quantum companies will fail, lots of small, four, ten, twenty person Quantum companies that are late to the game will fail. There's two hundred and something Quantum companies in the world. In three years there may be four. Every industry goes through this, every industry goes through a consolidation stage. The irony of all of this is there are only four thousand people in the world or so who are Quantum experts. So those four thousand people are going to appear or reappear in the one, two, or three Quantum companies that are going to win this. Canada has a chance of having one, two, or three of those companies based in Canada, from a hardware perspective. The Canadian government should take an opportunity here to say, ‘this is a pivotal point in the history of computer science. This is a once in a lifetime change in computer science. This will basically set the state for the next hundred years of computer science” This really isn't in any debate at all.

We shouldn't allow the foreign country to take it away from us. We have it here, as we have with lots of other things. We should double down on supporting those companies. Even if those companies eventually become two companies or one company, it doesn't really matter. It's a fact that people live here and like to work here and so forth. That's important.

The second thing that I think they don't fully understand is that a quantum computer is not useful unless it has quantum applications running on it. Quantum applications are typically written by Master’s, PhD type people. Ten people can write an application to do some quantum chemistry that could change the nature of climate change, for example.

Small numbers of people with this amount of power can change the whole world. We have an opportunity here to create what happened in Silicon Valley 30 years ago in Canada. I'm actually more excited about the surround, the quantum companies in the application space, around the quantum computers that we have here, more than that than the quantum computers themselves. Quantum computers themselves are, to a large extent, a solved problem. We will solve that problem. It is going fast. You need these applications people to change medicine, material science and so forth. We kind of ignore that knock-on effect. For example, I often get politicians saying that, “in Photonic you're creating 300 jobs. 300 jobs isn't interesting.” Well, it's interesting to me because these are super PhDs. These are the best in the world, people. They create five jobs for every job I put in here, but more importantly, the application surround around this is going to create thousands of jobs and change the GDP of a nation, which is not necessarily a function of the number of jobs that you have.

Especially with AI, people are understanding that actually fewer jobs can lead to higher GDP growth. So if your focus is on GDP growth, it's a false logic to believe that actually equals jobs. It did in the 18th century, but it's not going to do it in the 21st century. Let's get a grip here, people.

KB: Is there anything that I should have asked you that I didn't ask you?

PT: Canada is historically risk averse and has a history of losing commercialization south of the border.

To some extent, we've made that easier for them because there are three hardware companies, and this is not 300 now, four if you include Anyon, and so the winners have kind of picked themselves. They've fought in the trenches here to get to this, and the Americans see it. So fund them.If you want to, pick any two out of four, three out of four, one out of four, and then fund it. So as far as I'm concerned, as a Canadian, you're going to have to fund it. And the benefit is huge to Canada.

Those companies, by the way, are fiercely Canadian. So this view that we are in it to sell to the highest bidder couldn't be further from the truth. We are companies, we do have cap tables, we do have investors, we are subject to capital flows. But speaking for Photonic, Photonic is two thirds owned by Canadians, and it's fiercely Canadian. We’re banging the table going, “I want to stay Canadian. Give me a chance to stay Canadian.”

It's not a given that you form a company and just sell it. That is not a given at all. If you can form a company and have sufficient revenues, and have revenues with high gross margins, like Johnson & Johnson, which is a private company, you don't want 1) to sell it or 2) go public. There's no need to raise capital once you get past this kind of stage that you're in. To be a great company operating is actually a value thing that people do not understand. Often companies run into cash issues, so they are forced to take capital from other places. And of course the capital the Americans can bring to bear is high. The irony of all of this is the capital for quantum is a fraction of the capital for AI. A fraction. If you're an AI company and my PhD was in AI, so I can speak with a certain amount of knowledge, your cost base to run your LLMs is enormous. Absolutely enormous. So in quantum, the cost base is actually small. So it is a very strong high gross margin business, which is good. So you want high gross margin businesses that are actually creating opportunities for other businesses, which will also be high gross margin businesses to enter GDP. That's basically what you want to do as a society.

Interview with Allison Schwartz of D-Wave

KB: Tell me a little bit about yourself, about D-Wave, and about what you're trying to achieve.

AS: I’m Allison Schwartz, I'm the head of global government relations and public affairs at D-Wave. I've been with the company now for almost five years. D-Wave is a full-stack quantum computing company, which means we have the computer, hardware, real-time cloud access, support services, professional services, and all. We have right now a quantum annealer, which is in the cloud today, solving real-world problems. We are also building a gate model system, and we are using it through superconducting chip fabrication.

I mentioned the two different systems, not to get into the science behind them, but to highlight the fact that quantum computing is not monolithic. There are different modalities or different qubit architectures. Each of these are advancing on a different timeline, and each of these have different strengths and weaknesses and capabilities.

Annealers are best for optimization problems, which are ubiquitous across industry. That's what's commercially available to now, and we're in our R&D phase for the gate model systems.

KB: Is a quantum annealer a hybrid system, or is it fully quantum-based?

AS: It's 100% quantum. We do have hybrid solvers that you can submit your problem in, and you can ask it to go to the QPU only and have the quantum annealer take care of all of it. But a lot of the problems actually don't need quantum to solve all of it. A hybrid solver might actually be a better solution. If there is the part of the problem that HPC is really good for: a hybrid solver will break the problem down, and the end user doesn't need to do anything. They submit it into the problem, and then our hybrid solver will break the problem down, send part of it to an HPC, send part of it to a QPU, and then give you back the answer, and most often sub-second response times.

KB: It's pretty clear that Canada is a leader in quantum and that we have an incumbent advantage in the space. What is Canada doing well, and where are the gaps?

AS: Canada had a “leader” position in the academic space, especially on the hardware side. D-Wave came out of UBC, and so a lot of our technology came out from there. A lot of our employees are from Canadian academic institutions, and we still have our research center of excellence out of the Vancouver area in Burnaby.

What Canada is doing right is investing in the hardware and seeing the value of quantum. Where Canada can grow is on the application layer of seeing where these actually can do useful things. We need both. We need both the investments in the hardware advancements as well as the software, and we need investments in all the different modalities that are out there. You're seeing investments in, let's say, the EU and the UK that don't have as much of the homegrown technology. So they are investing in Canadian companies, coming over and building those applications. It's about the end use for the technology. How do I solve my problem, whether it be transportation or personalized medicine or emergency response. Where Canada is also doing really well, and I know it's going to surprise people: immigration. It's actually easier, if we have a talent that we need to bring into our Canadian entity, to bring them there versus our US entity. In the US it’s harder because of the immigration laws to bring somebody in there who isn't already a US citizen.

KB: With respect to quantum strategies and comparing Canada, the US, and the UK, how are others doing it differently?

AS: The first difference between the UK and the US is in taking an explicitly inclusive approach on quantum, including quantum annealers as well as gate model systems. This was actually the UK approach to their SPARQ program. They talked about having to train on annealers as well as gate model systems. That inclusivity has also moved into the EU quantum programs. It's also in the Canadian quantum strategy. It's not in the US one that was first passed, but it is in the legislation since to reauthorize the national quantum strategy. That inclusivity language has been folded into it. The problem and the gaps happen in the implementation side. So you have these national quantum strategies, but you have quantum programs that are implementing them, and there's a gap between what the strategy says and what these programs implement. There is not a good oversight or course correction when those gaps occur.

The UK actually has instituted programs to be inclusive of all technologies. The EU just did a call for air traffic management that looked at quantum sensing and quantum computing and explicitly included quantum annealing technology as high yield, low risk, short term technology. The Canadian programs, unfortunately, if you look at the last two or three that relate to compute are for gate model systems only.

D-Wave is the third largest quantum patent owner in the world behind Google and IBM, the largest in Canada, and we are explicitly excluded from their calls. So that is a concern. So, how are we going to course correct that given that you have a program implementation that doesn't actually align with the roadmap that just came out of ISED or the national quantum strategy that's come out of ISED?

KB: D-Wave was explicitly excluded from the DARPA QBI program as a result of the focus on gate model systems, and you mentioned being the third largest owner of patents in the world after IBM and Google, which is of obvious relevance to the ongoing policy debate around IP ownership and sovereignty. Tell me a little bit about the process of trying to actually get the funding you need to build this company, balancing that against IP strategy and retention.

AS: Let’s start with the QBI program and then I can get into the investment side of things. Originally, the DARPA program was called Utility Scale. It was just about getting a quantum computer to Utility Scale and D-Wave did submit an application. They had identified four companies that they were interested in, all of them gate model only, and two of them made it onto phase two: PsiQuantum and Microsoft. Then they came in with the Quantum Benchmarking Initiative, the QBI, which was an even smaller subset and that one they actually designated it for gate model only.

While we are building both annlears and gate-model systems, we did not submit our gate model program and the reason is the Quantum Benchmarking initiative had a six-year timeline and we believed that our gate model program would get to Utility Scale, but we weren't sure if we could get it in the six-year time frame. We knew we could get there with our annealer, even though it was technically out of scope. Under the Utility Scale the annealer should have been in scope, but DARPA had determined they only wanted gate models, so when they shortened it, they focused it there. There are 18 or 19 companies that got through the QBI program Phase I. DARPA is assuming a majority of them won’t make it to the next phase. So they do have a bit of weeding out and people who have great ideas but can't implement them in the timeline they need or have problems scaling.

As for the investment piece, because D-Wave has most of our commercial success is in the annealer and we are building the gate model system, but it is in the R&D phase. We're not building a NISQ-era system. We're going for fully fault tolerant from the beginning. We know that there's a longer runway for that. We are not beholden to government dollars. We actually have more commercial customers, so our commercial success is because we have real-world applications in production today. We're optimizing the telecommunications network in Japan with NTT DOCOMO. We're optimizing automobile manufacturing of Ford Otasan in Turkey. We're optimizing some operations in Canada for Pattison food groups, the Save On Foods in Western Canada. They're looking at a few different grocery optimizations including e-commerce delivery staff scheduling and others. So we have real-world successes pinging our systems in real time. We went public in SPAC in 2022. You're right, a lot of the investment did come from VC investors, but we did have PSP as one of our major investors in Canada. They have since sold their shares, which was announced publicly, but they were bought in fully into supporting D-Wave and our technology advancements and moving forward. But when we went public, we went public through our SPAC. Our parent company is a US entity, although we still have the Quantum Center of Excellence in Canada in Burnaby where a majority of our employees are. And that's where our products and services are being developed.

KB: Recently we have seen the G7 and the PMO mention Quantum. This conference and recent NATO recognition of quantum as a key emerging defense technology all show intention to push on quantum tech. What concrete steps need to happen now that the intention has been expressed?

AS: There are a few concrete steps. One is course-correcting the implementation of current programs, making sure they're inclusive of all the technologies, and that really has to be done almost immediately. There's been a second call now that has come out that has not been inclusive and there aren't that many calls. If you don't course correct quickly, you're just going to widen that gap.

There's been a conversation for many years now about creating a quantum sandbox. It's specifically focused on near term applications, identifying a need, identifying applications, whether they're quantum or quantum hybrid, and solving a problem. Supply chain, for example. Can we optimize ports? Can we optimize railways? Can we optimize airports? How do we move things faster? How do we move things better? There was an application in Tokyo which optimized waste collection and it reduced the amount of CO2 emissions by 60% because they could collect waste better using fewer trucks on the road.

The government and Prime Minister Carney have mentioned housing being a critical priority. Well, part of that is a manufacturing problem. How do you manufacture things? Optimization of manufacturing is right in the wheelhouse of D-Wave. We’re already optimizing manufacturing in a variety of different industries. We could do it for the housing industry too.

Supplies have to come to the ports. I mentioned port logistics, that's another area. So there's ways in which you can identify those policy initiatives, put it into a quantum sandbox and build those applications. The UK has done this and they put a timeline on those of 18 months or less. You can do that here in Canada. We can make utilizing today's technology an actual priority, and not just studying it and thinking about what it could be one day in the future. That is a concrete step that could be done through Innovation Solutions Canada through partnerships with digital or the quantum algorithms institute. We're actually working with Staque and Verge through a digital funded project on autonomy for farming equipment in Canada. That autonomy is not just for farms. There is a national defense area. There is a variety of transport that's looking at it, too. So utilizing current programs or current funding structures and creating a quantum sandbox program that is dedicated to near term applications.

KB: It sounds like this is something that you've pitched before. What's the response when you get into this with the government in particular?

AS: “Wonderful idea. We'll get to it.” I think it's time. I've been with the company now for almost five years. I think it's time we got to it.

I testified before Parliament almost four years ago. I pitched the idea of a quantum sandbox and everybody thought it was a great idea. It was actually a recommendation from Parliament to the administration and we're still waiting for it to be implemented.

What it does is actually from an investment perspective is that by building these applications and building the demos, it actually de-risks the technology, which brings more investment in. The quantum sandbox program, it's not just about solving a problem. It's also de-risking the investment in this emerging technology and demonstrating that it actually can do something useful. Through that, we can know a company is one we can invest in, or that a startup is one that we want to navigate with. These technologies can work with companies that are coming out of Creative Destruction Lab as well as the hardware and software that come in. Which Canada has a lot of. Right.

KB: We’re in the midst of geopolitical chaos currently, which is part of the impetus behind a lot of the conversation around technological sovereignty. Has that shift in policy impacted D-Wave’s IP strategy?

AS: It has not. The way we look at this is the IP we need is what we need to advance and scale our systems. We just released our sixth generation system. It has 4400+ qubits. It's more performant than our fifth generation system, which had 5000 qubits. It's not about the number of qubits. It's actually about how well they work together. What is the IP we need to continue to make that move forward? That's what we're looking at. Where is the talent that we need in order to help advance that? We're already looking and identifying ways to build Advantage 3, which is going to be yet another exponential growth in our hardware, but also the software side. So it's not an IP strategy. It's more about talent. It's what we need from building it. We're going to need the right people, and if those people are in Canada, fantastic, but if those people happen to be in the UK, in the US and other places, we're going to have to go where the talent is.

KB: Can you give me a sense of scale for the optimization problems that you're looking at? Can you compare what it would take to solve that using a classical method to what you're able to do with your annealing systems?

AS: So it's hard to answer it from that perspective because there are some optimization problem sets that actually still are better on classical. We worked with the Australian army on autonomous vehicles for emergency response. When you looked at five trucks going from two depots to seven places, a classical computer could solve that. The minute it got to scale of an actual problem, you had seven depots, 300 trucks, etc., by the time you got the solution from a classical solver, it was too late for an emergency response, or it never got to that answer. Our hybrid solver could get there.

It's time to find the solution that you're looking at. If you are trying to optimize wildfire remediation and response, ERDC, which is an army research lab down in Biloxi, Mississippi, identified an optimized fuel break line. How do you actually do the fuel breaks to prevent those cascading wildfires?

It's not that you couldn't do fuel breaks before, it's not that classical can't do it, but we were able to optimize it so that they could tear down 60% less trees. They're also looking at cascading grid failures, which was important with Spain and Portugal recently. As you're looking at the wildfire response in BC that happens every year, it's not like we have an application that we can plug and play. You can't call us during next year’s fire season. The problem is that we have to build those applications now, so we can test it out next year. So it's less about the scale of what a classical do versus a quantum, it’s more about time to solution, and we do a lot of benchmarking in the beginning to identify whether or not we even need quantum for the solution that you're looking at..

KB: What should I have asked you but didn't?

AS: I guess the question you should ask is why people aren't inclusive, why the government programs are not inclusive of all the different technologies. A lot of that is that because quantum is scary, because quantum is so niche, they have a lot of folks that have 40 years worth of academic background that is forcing them to one viewpoint and they're not asking, “what are we not thinking about?” They kind of have that preconceived bias that's going in. There are some governments around the globe that have what's called a 13th man perspective. If 12 people agree in a room, the 13th man has to play devil's advocate. Unfortunately in the quantum programs, whether it's in the US, Canada or others, they don't ask “what if we're wrong.” China, on the other hand, has an all-inclusive approach of just going after everything. When I continue to talk to governments around the globe about what they need to engage and where they can look, when I say, “but what if you're wrong?”, they don't understand the question. So they really have to have a “what if we're wrong” mentality and do GAP analysis. If these are all the technologies that are out there and we've implemented programs, where are the gaps for talent? Where do we need to invest in talent in those areas? Is it applications? Is it annealing? Is it hardware? Is it sensing? Is it networking? There's a variety of these different niches, but we're not doing gap analysis and implementing programs from that.

Last week I attended QUANTUM NOW | ICI QUANTIQUE, Quantum Industry Canada’s tech expo during the international year of quantum. Given recent recognition by the PM and his G7 counterparts and NATO of the importance of quantum technologies, the timing could not have been better.

Canada has an incumbent advantage in this space, and as a result is in a position to lead the world through the disruption that quantum technologies promise in computing and beyond, and while there is still work to be done, the ultimate beneficiaries of the technology are being decided today. This conference was a well-delivered and focused forum toward establishing the roadmap we must adopt to ensure that we maintain and cement our lead.

In this post, I compile a set of reflections, lessons, and observations from the conference in advance of outlining a more concrete roadmap that Canada will need to adopt in order to secure its position as the global leader in quantum technology development.

Lessons from Sherbrooke

The conference started with a tour of the various moving parts of the quantum valley in Sherbrooke, QC, one of the oldest hubs of quantum activity in Canada.

In a previous interview with Lisa Lambert, we touched on the circumstances that led to Canadian quantum leadership. In Waterloo, it began with an eye-watering philanthropic donation of one third of the personal wealth of Mike Lazaridis at the time that led to the founding of the Perimeter Institute. In Sherbrooke, the path was different.

In 1980, the university decided that, being a small faculty, they could achieve more by specializing and concentrating their resources into one area than they could by taking a broader approach, and they went all-in on quantum materials research.

The government of Quebec soon got involved, and has poured money into it over the years, money that has been paid back many-fold. The concentration of expertise paid off in the long run and became a self-reinforcing thing. Concentration of world-class research talent and a focus on dedicated infrastructure led in turn to attracting more talent, to the point that Sherbrooke became a globally known hub of quantum research. Distriq, 3IT, and Institut Quantique together form a pipeline from lab to market that has already delivered commercial value that far exceeds the cost of building them. Institut Quantique boasts more than 600 graduates, more than 200 HQP positions created by spinout companies, and $500M in follow on investment, about 14x the original grant.

So deeply is quantum embedded in the region’s identity that even one of the buildings we visited was architecturally inspired by quantum machinery, and there is a real sense of excitement from all involved over what feels like the end stage of a process started almost 50 years ago.

Canada can learn important lessons from Sherbrooke.

Lean into regional concentration

In Canada, it has been pointed out by many that we tend to spread out support for innovation, sprinkling resources across far too many priority areas, both geographically and across sectors, to move the needle in any one. This comes from a place of good intention: fairness is a deeply rooted Canadian value.

However, if you look at innovation ecosystems that are successful, you will find that they all have in common a strong regional concentration of talent and capital. This is a natural and, in my view, largely unavoidable emergent property of the way innovation reinforces itself, as talent and capital come together to create more of both over time, just as Sherbrooke’s focus on quantum excellence reinforced that excellence.

One of the reasons that it has been difficult for Canada to establish technology-focused innovation hubs generally is that larger ecosystems tend to feed off smaller ones. Canada’s tech hubs are in many cases just feeder hubs for San Francisco and Boston. Once established, these ecosystems tend to be very difficult to dislodge, making first mover advantage critical.

What is sometimes missed in discussions around how to manage this, however, is that this effect is not just geographic, it is sectoral.

Sherbrooke (and Waterloo) prove the value of leaning into regional specialization early and taking a deliberate approach and a calculated risk toward owning a particular space, a lesson that can be generalized to any emerging technology. For a tech hub like San Francisco to siphon talent away from Canada, there must be an established base of talent in the particular tech sector of interest, and the United States has very little quantum activity, relatively speaking, compared to Canada. Because of this, Canada has an opportunity to cement its quantum advantage and become the global attractor in quantum.

Admirable though the desire may be to ensure that investment is geographically fair and equitable, Sherbrooke’s example, suggest that we should reframe how we think about fairness where innovation is concerned. Instead of trying to ensure that everyone has a chance at getting support where they are, Sherbrooke’s example shows us that we should lean into regional specialization, focusing our efforts on creating conditions where small ecosystems can have disproportionate global impact. This is not incompatible with principles of equity and fairness, it just requires a reframing: fairness in this approach means ensuring that anyone who has something to contribute has the opportunity and mobility to participate in the growing hub that supports their specialization.

Reflections from the tech expo

QUANTUM NOW

In contrast to so many of the focused tech conferences I have been to in the past few months, QUANTUM NOW was a refreshing change of pace from the usual hype cycle. As a physicist that is not a quantum expert, I went into the conference expecting a lot of deep tech companies with huge amounts of research and technical uncertainty between them and commercial impact. Instead, I was greeted by scientists and engineers sharing unadorned stories of progress toward quantum technologies. Everything I saw felt grounded and real, and nobody used the phrase “game-changer”.

Many of the exhibitors have products on the market and are actually delivering commercial value today. Quantum technologies are no longer science fiction.

Quantum is not a monolith, and not all elements of quantum technology are at the same point relative to commercial impact. Quantum sensing broadly refers to the use of any quantum phenomena to sense elements of our physical world and probably represents the most advanced of the three areas. SB Quantum, for example, has developed sensitive magnetometers that are set to be launched into space by NASA next year, with the goal of mapping earth’s magnetosphere. Through that mission, the technology will provide an alternative to GPS for global navigation, using sensors developed here in Canada, and shows promise for defense applications as well as mining. Phantom Photonics has developed LIDAR sensors for long-range undersea sensing, promising the ability to sense submarines and monitor undersea cables, with applications in defense and arctic sovereignty, and is already engaged with NATO DIANA.

Quantum computing has been “10 years away” for about 25 years now, but it was clear from the exhibitors that value is already being created. Quantum computers with early commercial utility are already here.

To make the stage of development more concrete, one of the panels compared timelines for quantum computing to classical computing. Starting with the transistor in 1947, gate devices in 1958, processors and applications running on them in 1971, personal computers in 1981, and Microsoft OS and consumer computing taking off in 1990, classical computing proceeded on a series of 10-year breakthroughs that is now being mirrored by quantum development. Quantum is on a roughly similar if somewhat extended trajectory. The qubit was demonstrated in 1992; D-Wave Systems (also Canadian) released D-Wave One, now recognized as the first commercial quantum computer in 2011; IBM released their first circuit-based quantum computer in 2019, and recently demonstrated landmark error correction results suggesting that it will be possible to scale up to large qubit counts.

Today, this is where we stand, with commercial “mainframe” quantum computers a reality. Quantum Insider reports that 41 quantum computers were sold in 2024, at an average price of $19M. To put this in perspective, this represents a doubling of volume and a halving of price since 2021.

The general attitude among panelists toward the question of when we would have mainstream quantum computing was one of quiet confidence that it was just a matter of time and engineering, rather than a question of physical possibility. While quantum computers that meet the DARPA definition of “utility-scale operation” may still be 10 years away, that number seems to now be realistic. DARPA wants to do it in 8. A scaling race is underway, a race where Canadian companies have a significant headstart.

ICI QUANTIQUE

I want to take a moment to acknowledge the communications team that named this conference: QUANTUM NOW | ICI QUANTIQUE. For those who do not speak French, the French name is slightly different from the English one, translating as “QUANTUM HERE”. As came out clearly in almost every discussion in the conference and as I discussed above, there is a strong argument to be made for Canada to make a stand in and lead this space. Thanks to the aforementioned early investments in Sherbrooke and Waterloo, and emerging but still strong quantum hubs in Calgary and Vancouver, Canada has a disproportionate concentration of quantum talent. Of the people globally who are trained and able to contribute to pushing quantum science, engineering, and software development, many of them are in Canada. As such, Canada has an incumbent lead even over the United States. On the other hand, American and Chinese investment in quantum dwarfs our own. Canada has a very limited window of time in which to capitalize on this opportunity and establish the regional concentration of talent and capital needed to make Canada an attractor hub, rather than a feeder hub, of quantum technology and talent.

As with any emerging technology space, governments are still prime movers in quantum tech, spending $40-$50B per year. While VC is accelerating, it is still tiny compared to public sector spending, operating only at the margins. It is also clear that M&A activity is picking up, and many at the conference expressed an expectation of market consolidation into a few large companies over the next few years, a natural progression that signals an emerging market moving toward maturity.

This points to an unsurprising but important learning for policymakers: it is government spending, not private sector investment, that will decide who wins the race. If we want consolidation to occur in Canada’s favor, we need to set our companies up to direct it.

Just like the name of the conference that can only be fully appreciated in both official languages together, a coordinated approach that includes both anglophone and francophone Canada is necessary for success. Both have established internationally renowned centres of quantum excellence in research, centres that are now transitioning into operationalizing this research through companies that are advanced well past the startup stage, companies that are ready to anchor Canada as the global leader in quantum tech—if we stay the course, and unite behind a pan-Canadian push to be the first country to fully realize the potential of quantum tech.

Quantum is happening now with or without us, but if we want quantum to happen ici, we need to double down on the commitment, and play to our strengths - strengths which require coordination across both anglophone and francophone Canada.

As someone with a deep appreciation for effective communication: chef’s kiss.

What’s next?

The vast majority of value creation from quantum has yet to be unlocked. While the enabling hardware remains in development, just as with classical computing, it is the application layer in which most of the potential for value creation lies, potential that is only now entering the realm of possibility. You have probably heard of exponential speedups that are possible in cracking RSA public key encryption using quantum algorithms, but there are numerous potential applications where quantum computers can provide even quadratic speedups over classical approaches that have enormous potential for value creation. With only a few thousand people globally that are even qualified to build in the space, Canada has everything it needs to lead the way.

For now.

If you’re tracking the Canadian quantum space, you have likely heard about the DARPA QBI program and that four Canadian quantum companies are participating. While at the conference, I had the opportunity to interview leadership from several quantum companies, including some of those that are part of DARPA QBI, to better understand the implications for Canadian quantum. This post is already long enough and so the interviews and related commentary will wait for my next one, but the core message was clear and consistent: if Canada wants to remain relevant in quantum and cement its leadership position, the time to commit is right now.

I’m glad to see that Minister Solomon has this top of mind as well, discussing policy moves intended to ensure that Canadian quantum companies can stay Canadian. In my next post, I will use my interviews of the leadership of several of these companies to suggests concrete policy initiatives that can support this goal, and provide timelines over which they must happen if they are to matter.

This week I interviewed John Wilson, the CEO of Innovate Calgary and someone who has been involved in building the foundational elements of innovation ecosystems and commercializing emerging technologies in both the UK and in Canada.

My goal in this interview was to better understand Canada’s most developed example of a public-private, venture-philanthropic approach to investing in emerging technology, and to hear from someone with first-hand experience across multiple innovation ecosystems what needs to happen to address Canada’s perennial challenges with the process.

The history that John shares is fascinating, and his commentary makes clear that innovation is a deeply interconnected process that requires patience, risk tolerance, and a cohesive approach to support that recognizes that everything—money, intellectual property, and acceleration, in John’s words—must come together into a cohesive framework for any of it to work.

Your email client will probably truncate this post. My key takeaways are presented at the end, so be sure to read the web version if you want to get the whole story. Many thanks to John for sharing his experience and insight.

Interviewer’s note: John Wilson and the Innovate Calgary team approved the final version of the section entitled “Interview with John Wilson” and had editorial input on that section, with the option to rephrase and expand on the ideas discussed in the interview without changing or removing any intended meaning. The key takeaways presented at the end are my own commentary, and do not necessarily represent the views of Innovate Calgary.

Interview with John Wilson

KB: As I understand it, you've been involved in supporting emerging technologies in various roles for most of your career, starting in the UK and now in Calgary. Tell me a little bit about yourself and give readers a brief history of your experiences.

JW: As you know, I'm a self-declared techie. In the UK system in the eighties or nineties, you could end up with a PhD by the age of 24, which might be a bit young, really, but you can barrel through the system, and I did. Amazing, of course. I spent five years as a postdoc, and I took the opportunity to go to the US.

I went to Virginia Tech, great place, and then came back to the UK. I've always been looking at academia. I love research. I had the opportunity to join Unilever, a large multinational that had over 200,000 employees when I joined. I could do academic research, I had– fully funded PhD students, as well as doing commercial work and getting involved in product launches. So that was really excellent.

And then, I had a good friend, a mathematician, who wrote a paper that caused some interest. It was a mathematical paper, but it suggested that a different type of imaging was possible where you could get depth from images using fairly traditional optics. And that caused enough of a stir that we agreed that I would leave my job, become the CEO, and he didn’t need to leave his job.

We had no problem raising finance, and this is just based on a paper. We put a team together, ten, twelve people most of the time, set up shop, worked for five years. We had some wins and some losses. The technology was in a very competitive space. It wasn't going to be cost competitive. The company started design work, which we were good with–lots of good software people. At that point, my utility was questionable. So, I had the opportunity to become the CEO of another company. I said, “no, thanks. I'll go to Oxford University.” So there we are. I'm giving myself away.

Oxford is… look, they like all the gowns and the history and such like, but they are very progressive. We're going to touch on that in a minute. But they, as well as a few other universities, were really looking to the future. So, again, a really privileged position for me to be there at a key time in history.

I came to Canada nearly fourteen years ago. I went to Brock in Niagara. It's a different project. They were looking to rebuild an office. Brock has over 30,000 students, smaller research. So, this was much more about student entrepreneurship. I enjoyed that. I had to be dragged out of Niagara. It's a fantastic place.

But UCalgary came calling. As you said, I've been here nine years. I think we've built something quite interesting.

KB: It's my understanding that the university-attached seed fund and emerging technology support ecosystem in the UK is at least a decade, if not more, ahead of Canada's, and it's evolved quite a bit. I'd love to hear about your experience building those seed funds, and what lessons you took away from that. What can be generalized to Canada, and what can't?

JW: Some of the history has been told to me. I wasn't there at the beginning, and we don't want to go too far back. I think you can go back to the eighties. It's the start of a much heavier involvement in university research coming out of the US, the two big areas are health and digital, biotechnology was just taken off at the beginning of the 1980s, and, of course, digital with the silicon revolution.

These are things today that we see as anchors for all of our technology, health and digital, essentially. They were really starting in the late seventies, early eighties, with universities playing a role. At the same time, the US changed some of its rules around managing IP, devolving from a central management process, which was seen to be cumbersome and removed, and using the universities more as a proxy. [[Interviewer’s note: here John is referring to the Bayh-Dole Act]].

It's one of the one of the big debates that our industry has to this day: what was it that sparked the US industry that we see today? Was it the technology boom, or was it the change in the regulations? The answer is probably “yes”, Kyle. Oxford Innovation Company, which then was called Isis Innovation, was set up in the early eighties as a direct response to this. They could see that the Brits had to do something to compete.

So, they set this up in the eighties, and a few other universities did the same. And then in the late eighties, early nineties, it was apparent that access to capital was much easier in the US, and we just didn't have those early-stage capital access vehicles in the UK. Ian Cooksey was on the board. He was an early UK VC and had one of the UK's First VC funds. He was also connected with Oxford University. He helped set up Isis and helped to lobby the government during the nineties for seed funds for universities.

In the late nineties we had a new Labour government, and they introduced something called the University Challenge Seed Fund in '99. I don't know how they came up with it, but their model was that the government would put in some money, two foundations put in some money: the Gatsby Charitable Foundation and the Wellcome Trust. Any university that wanted to dip into this pot, and the pot was £40,000,000 GBP in the first instance, would need to put in 25% and then you could have up to 10% of the pot, I think. Oxford started with £4,000,000, and we started UCSF at Oxford in the early 2000s. Cambridge did the same, as did Imperial and UCL, all the really big universities that you could argue were trading on their research, but they're all being progressive and now looking to do more in the innovation space.

Regarding the investment process, I think it was up to £300k GBP right from the beginning. So let's call it half a million CAD. It looked a bit like a Dragon's Den process before we had Dragon's Den. It was mostly projects coming from the university TTOs. It could already be a company, but you didn't have to be a company at that point.

I would work with professors, build the business case for them, and I would make the pitch. were fixed rounds, and we would get a ten-minute pitch.

You would say yes or no that day. The success rate, I think, was deliberately designed to be around the 50% mark. The sweet spot is you don't want it to be just a foregone conclusion, but just because of the effort we're putting into it, you can't have a 10% success rate. 50-60% is not bad. For a decade, and for the time that I was there, we're making investments where the assets are accumulating, but there's no return of capital at this point. But the managing director of Isis Innovation at the time was Tom Hockaday, and every speech I saw him give, he would lead, or it would be a headline, that the biggest impact on the culture of innovation at Oxford University was the UCSF.

For a professor to get on the train in Oxford and go down into the city and ask for investment was very unusual. One or two always did it, but the vast majority were not going to do that. But to come to a university affiliated agent and be part of a process, and for me, or someone like me, that is much easier.

It was an amazing cultural success, so that was the early win. But then just after I left, Oxford Science Innovation set up what eventually became or very quickly became a £600,000,000 GBP fund just to invest in innovations coming out of Oxford. One university!

But UCSF was not a success in every university. For many it was, but some universities started, and it didn't really take off. I think what that tells us is that money is great, but it's never all about money. If you don't have the other bits and pieces around it, then it doesn't work. So that's a really key learning, I think.

KB: It sounds like that first forty million GBP fund was set up as a charitable public private partnership. Is that right?

JW: It's not for profit. They're evergreen funds. All the proceeds were turned back to the fund. This fund took about fourteen years to become evergreen. The aim is to fill this gap, and for it to be accepted that you could lose money, and that's still okay.

KB: UCeed is probably the most advanced Canadian example of the venture-philanthropic, public-private partnership model for supporting emerging tech. Tell me about Innovate Calgary and UCeed, and the process that led you from the knowledge you brought from the UK to that structure.

JW: The Innovate Calgary (IC) you see today, is not a replica of what then was Isis Innovation, what today is Oxford University Innovation, it is based on the understanding that you've got to have enough to do the whole job and not just bits of the job. Otherwise, you fail. We sometimes use the phrase “one stop shop”. I don't know whether that's a great phrase, but it conveys that we are trying to join things up. Innovate Calgary was set up in the eighties, so we will be 40 years old next year.

IC is an economic development agent for intellectual assets. That's my nice concise phrase.. “Intellectual assets” I prefer to “intellectual property”. I just think it's a broader, more general term. We do make money sometimes. We do not reliably make money. So, again, we're an economic development agent. No one's investing in Innovate Calgary. Banks aren't loaning us money. That's important to understand: we're an agent, not a company trying to make money.

The company is divided into three main groups: IP, Investment and Acceleration.. The acceleration bit you could break in two: physical assets and human assets. ,.

So, UCeed is a program, five years old, this year. It supports all three pillars of the university: teaching, research, and community. The capital comes either directly as donations from individuals or it comes indirectly, from foundations. Donors have made a donation to a foundation, and then the foundation wants to work with us and set up a UCeed fund. So the money comes to the university in those two ways, and then we create a fund, each having its own term of reference.

We have seven funds in total at the moment. It would be complicated if we had to then run a separate program for each seven, so we don't do that, we have groups: health, science and engineering, and social. Whatever comes in, we manage it in one of these three groups. We’re still learning, but we think that allows us to take on any donation, any grant, without it causing us more pain in just the management cost of all of this.

Our seven funds: We have three in the health space: we have child health, which was our first, general health, and neuroscience. We have two in the engineering/science group, they are energy and engineering and science (fund launching Fall 2025). We have one social fund, and we have a student fund as well. The student fund can invest in any sector, and it's a class in the Haskayne School of Business, a class of 15. It's a two year class open to any student at the University of Calgary, in any program. These guys and girls are leaving university having run a private equity fund for two years. So it's a pretty interesting course. We are trying, and again still learning, to try and get that fund integrated with all of our other funds. I think for me, the holy grail is to have the student fund running across everything we do and somehow involved in the management of the whole process. But one thing at a time.

In terms of key components of the seed fund, firstly, it is an open process, we often have outside observers. It's a dragon's den type process borrowing from UCSF. There's no privileged access to any of our advisers. We have an advisory group of people for every investment round.

The due diligence we do is public, we give it to the company. We don't market the due diligence or such like. We do the DD package, we give it to the company and then we say, “look, there you are. If you want to burn that, you can. But if you want to use that, you can as well, and that's fine.”

Then, of course, we have to manage conflict of interest because many of our advisers are active elsewhere and could be conflicted. They may have a connection with the company, so they recuse themselves, and we have a process for really trying to make sure we don't make any mistakes. We have amazing staff, and we take this very seriously. We're dealing with highly respected organizations, the Alberta Children's Hospital Foundation for example, so lots of effort goes into managing conflict of interest.

There is also a community dimension to UCeed. Each of our advisory committees is about seven people, 4 committees, so that's twenty to thirty people engaged. It's great community engagement. Student involvement is massively important. I'd like to have more students involved.

Then lastly, our relationship as the innovation office with the advancement office. The head of advancement at UCalgary is Andrea Morris. She's amazing, of course. UCeed doesn't exist without her. As an aside, when we introduced this, the past CEO, who I remained in touch with came to me and said, “well, congratulations, John and Andrea. You've succeeded in doing what the last five CEOs of Innovate Calgary have been trying to do.”

KB: Is the pitch to the donors typically around community impact and building something, or do they seek to seed companies to later directly invest? What's the typical connection between the donors and the outcomes?

JW: As always, there's more than one motivation. Foundations provide significant funding for research, the Alberta Children's Hospital Foundation (ACHF) is a significant funder of research at UCalgary, and most other big universities have these stories as well. So tens of millions every year go into really important projects. And then five years later, where does it go?

So, there is a need to translate the research. The foundation could do that, but do they have the infrastructure? We spent over a year with ACHF, and they were our first group. Andrea and I had regular meetings with their board, with their senior executives, explaining how we would safeguard them as well as introduce this really rigorous process which was going to be the most effective.

Again, you know, I have to thank everyone. They said “yes, and you guys look like you are capable and a good partner.” So, they trusted us, and we're still there. Great relationships.

We've had at least one other foundation come in, but then we have individuals as well. For energy and science and engineering, two people in the city, amazing individuals, heard about us, came as an observer, and watched the process. By then, they'd seen some of the investments. Maybe they'd even read some of the due diligence we'd done through various parties. They could see that this was adding value. One was a $3,000,000 gift, one was a $5,000,000 gift, and their motivations are giving back.

KB: My understanding is that UCeed does very detailed due diligence, and unlike most investment funds, you give it to the company. What can you share about what goes into the process and what goes into the DD package?

JW: I'll oversimplify first because I think this is the most important bit, and my colleagues will have to forgive me because I'll come back and praise them later. The real magic here is that we do it at all for this stage investment. So, we've never really put a price on what this DD package could be, but it could be $50,000. It could be more than that.

Who in the private sector does $50,000 worth of DD for a $50,000 investment? You can see where this is going. You'd have to be crazy, wouldn't you? So, there's the secret. No one in the private sector is going to do that, but if we look at our mandate and say, “well, this is what we're doing, folks,” then we do that. We do these outsized DD packages.

Coming back to this one stop shop philosophy. We have IP, we have marketing, we have finance. There are other organizations who could write this package, but I'm going to say that we are as good as anyone, and by the time you've done 50 or more, you'd like to think you're getting a bit better. So they are good packages. But the real secret is that we do it at all.

Since we met about a month ago in Ottawa, three companies have come to me, unsolicited, and told me how useful the DD package was for them. I'm going to say unsolicited unless somehow, I am teasing it out with them without knowing it. I met John Wong from Fluid Biomed, and he started telling me about how he was using this package. I met Steve Edgett at Genomadix, and then a week later, I was in Montreal when I met Jeff Bergthorson from Altiro Energy,. The investment is important, but a third-party review, if it's of a certain quality, is really valuable to our early-stage companies. We're tracking that, and we're somewhere between 10x and 15x between twelve months of UCeed investment. Jacob Johnson from Innovosource has much better stats for the US because they’ve got over 50 of these funds running, it's about 15x, in the US. We're also hearing now of how the DD is being useful for them for next round investment.

KB: As you say, there's no way that the for profit fund could pay for that. That's just cutting into IRR, whereas the nonprofit approach gives you some leeway.

JW: Quite simply, there are private early-stage actors, like us. Often, they will be candid and say that when they do these very small investments, they are business development, and very little formal DD is done. They will make their early first investment. They'll use that to look at the company, and the real value then is their follow-on rights and their later investments. But this is business development for them. Don't spend 50k looking at it. Make a small early investment and let the business develop. We could do that and save ourselves a lot of money, but then we wouldn't be fulfilling our mandate of leveraging all the others. I'm not criticizing those companies that don't. They are for profit companies. That's what they do. But you can see that we have this discrete role that we are playing, and that makes us feel comfortable that we are filling a niche.

KB: You mentioned seven funds, and it sounds like these funds are mostly chosen in response to the desires of the donors that seeded them. Are you going out to solicit specific donations in specific sectors, or is it mostly driven by pull from the donors themselves?

JW: Our first fund was with the ACHF, there was already a deep relationship with the university, we thought, and they’re such a respected organization..

The student fund was important to us. I think just as COVID was breaking out, I walked into the dean of the business school's office. I said to my colleagues at the time, “I don't know why this guy should say yes, because there are other things on his mind right now.” But you have to be optimistic, so I go in, and I'm an unknown person to him, he knows I'm a tech transfer person, so that means I'm a troublemaker. I walk into his office just as COVID is breaking out and say, “Jim, I've got an idea for you. Why don't we set up a private equity fund in your faculty?” And inside I was laughing, I was expecting a bit of a defenestration. And he said, “yes and let's do it as quickly as possible.” Every now and then, Kyle, you come across these great characters that just unlock the door. Now we have a new dean and she's brilliant as well.

Sorry for the anecdote, but it amuses me. The Social fund, absolutely a must, and one of the trickier ones to set up because “social innovation” still means different things to different people, so it has a communication challenge, and it has a challenge of keeping the broad church together. But and I think it's true in many universities, if I were to ask every faculty member which fund do they most associate with, then I expect the top fund to be the social fund. We have partnered with the United Way, Calgary and area, so again, more relationship building, but now with three or four funds in, we've more stories to tell.

And then the private donors came in a bit later. Energy, this is Calgary. Engineering Science was absolutely fantastic. Really didn't spend long talking with this gentleman, and he's amazing.

So, I think they've chosen themselves, and I think they will in most places. I can't imagine what fund would come to us that we couldn't do, but maybe, I don't know. I think just because of our relationships, they've naturally emerged. I think they've been significantly driven by us, but not always.

Maybe in the future, there could be an Ag fund. Ag is another one of these broad-church areas. Trying to get two people to agree what Ag means is challenging.

And then there is the issue of a follow-on fund. We're five years in. We've made 80 investments. Is there a role for us to play in the next round, which should be a completely different fund? It wouldn't be UCeed. It would be an investment fund.

Should we do that, or should we let someone else do that? And that's one of those discussions that is ongoing.

KB: You've been at IC for about nine years. How does the early stage innovation ecosystem in Canada broadly, and Alberta specifically, compare to the one that you helped build in The UK? Where is Canada doing a good job? Where are the gaps?

JW: There are definitely leading actors in both countries.

Waterloo has been the leader in entrepreneurship at least, but other things as well, and have been blazing a trail for decades. So, you have to take your hat off to Waterloo. But there are others as well. Over the last decade UCalgary has become a leader. But the leading UK universities started earlier. I have mentioned that the UK had the university channel seed fund in around 2000. Well, we don't have it here yet, and we're now 25 years on. I'm not saying we copy the British because there's clearly some things we shouldn't copy, but this really does seem like a good thing for very little money, so why don't we do it? I'm also not an expert on HEIF funding, this is the Higher Education Innovation Funding, which is central funding that the UK brought in, I think, 2001.

The importance of central funding is that if we ever wanted to do anything in a joined-up way in Canada, we will often hear, “oh, every university is different. It's very difficult to do joined-up things.”

I don't agree with that. Core funding to fill in the gaps allows very different universities to play together. So, if you wanted to have a seed fund for Canada, also having a HEIF type fund to allow all to fill in some gaps, whatever gaps there might be, that enables it. So twenty-five years ago, the UK had central funding and seed funding.

If you look at the UK and Canada, we clearly have very similar academic systems, a shared history. You can include the US as well of course.

The history is very devolved and market driven in terms of commercialization structure at UK, US, Canada, and other British derived universities. But in other parts of the world, the Universities are much more heavily connected to industry and were from the start. Japan, Germany, Switzerland and South Korea as examples.

If we come back to the UK and Canada, in a market driven approach, and in a country like Canada where we have a low receptor capacity for research, then we see poor innovation outcomes, and that's it. In the US, you've got significant receptor capacity, so the market operates. In the UK, you've got a bit less than the US and you've got less capital, which they tried to fill in, so they don't do quite as well as the US but better than Canada.

All this means that we should do something. I don't think that doing nothing is now an option. I've mentioned core ops, (HEIF-like funding), translational projects–(SBIR and STTR - like funding), seed funding, IP funding (Elevate IP for universities). So, if each of those programs was a hundred million each year, that'd be much bigger than anything we're doing now. Collectively, it'd be about 10% of SR&ED.

I couldn't get to the end of that without throwing SR&ED in. SR&ED is good, but it's a blunt instrument, and it's a large number. 10% of SR&ED would make all the world of difference in terms of our government led university innovation, and it wouldn't really change SR&ED at all.

KB: Within twelve months, you're seeing 10 or 15x in terms of follow-on investments. Recalling that in the UK it took fourteen years to become evergreen, with UCeed being about five years old, it probably hasn't quite hit that stage yet. What metrics are you collecting that correlate to progress toward this goal? How do you argue for the impact of something when the ultimate return is still years down the road?

JW: I inherited, for Innovate Calgary, a spreadsheet with something like 130-140 metrics, and they're all good, but we have reduced this number. The way in which we’ve approached this for a few years is that we've looked at the values that we have and the stories that we tell, and we have grouped our metrics into four categories: impact, finance, culture, and excellence in operations, and they are our values. Each group has 4, 5, or 6 top line metrics.

For impact, we're currently looking at our Pitchbook ranking, the number of startups, number of patents, the number of IP strategies.

Finance is income from various activities. We do make money, and although we accept that we can lose money, we also accept that we live in the real world, and the more we can make, the easier it makes things. So, we have a number of finance metrics. Impact and finance are the closest we have to our output metrics, and our other 2 groups are more like input metrics.

Culture metrics include: how many professors are engaged? How many community members are engaged. How many students. Culture is an engagement metric. I like to work with a quarter of our professors every year, and that's a big number, and this means innovation is cultural at the University of Calgary.

Excellence in operations is our last group. We sign over a thousand legal agreements every year. Much of my time, Kyle, is me reading and signing. So, it's a real machine, and we have to be good at doing this because it’s important and can be time consuming if you are not disciplined.

KB: It feels like the world generally, and Canada specifically, is at a significant political crossroads. What innovation and emerging-technology-related policy changes are you hoping to see from the new government?

JW: I'll make a comment first. I'm not sure if you consider yourself a policy person or that's your core. If you look at the history of university or public sector innovation and commercialization innovation starting at 1980, it was mostly led by lawyers in the early days, and you can see why: these are these are new types of legal agreements that we're doing, and you need to get some lawyers trained up to work out how to do this. So, if you look at the early tech transfer offices (TTOs), it was often a lawyer that was running them.

Then it started to get much more blended, and there's commercial people like me. What’s next? I would expect, in the next ten years, there to be more people in policy in leadership teams, in TTOs, because I think we know how the mechanics work. It doesn't mean we get it right every time, and we know what the problems are. What we're missing now is persuading the government to do it. Clearly people like me aren't very good at doing that kind of thing. So I'm really encouraging you, Kyle, not to leave this profession, I think, and to get policy involved in the innovation offices.

I've listed my priorities. There's five.

1. Core funding: Because core funding will allow every other program to operate. It’s the foundation course for a degree if you want to use that analogy. It will bring Canada together, understanding that Carlton might be at a very different place to UCalgary, to SFU, etc.

2. Gap funding - SBIR and STTR like funding. Our medical funding used to have things called POPs and that went away, which was translational funding. So, we actually have less than we used to.

3. Seed funding, clearly, and because I believe it's cultural as well as functional.

4. IP funding. ElevateIP could have a stream dedicated for University IP.

5. Finally, training and skills development, Lab To Market could evolve here.

If each of those were a hundred million dollars a year, that's half a billion, that's of order 10% of SR&ED.

I would write this up as one framework, so they all play into each other, and now you're spending half a billion each year. We should stop talking about tens of millions. Let's do it properly, in one program that puts all these together. That would be great for Canada.

KB: Is there anything that you wish I had asked that I didn't?

JW: Is there anything I've said that surprised you, Kyle?

KB: The biggest surprise was the relative scale at which these things are being done in Oxford versus Canada. The idea of close to a billion dollars of funding being pushed towards a single university, is mind-boggling just in terms of the implications for the amount and quality of emerging tech. Oxford isn't substantially larger than some of the large Canadian universities.

JW: It is quite large. It's not as big as U of T, I don't think, but I think it's bigger than UBC. So it is large, let's say it's a billion dollars a year. UCalgary is half a billion dollars a year in terms of research funding.

It is Oxford, of course. So, they've got a large amount of highly ranked research. But that doesn't explain just the complete difference in scale in this. That is because they've leaned into it, and the follow-on-fund is financially driven. That's why we should put together this framework, and we should allow for some failure, if it doesn't work in every university, the timing is wrong, but if it worked in a few of our universities, that would be worth the investment. I don't think we should do nothing right now. I think we've read enough papers on Canada about why we have a challenge innovating.

KB: That's a good point of comparison. The other surprise was getting a yes straight away from the university administration. You’ll have to teach me that trick.

JW: If this story was a movie, it the list of credits at the end would be a long list. You do need a lot of people to lean into this somehow. You don't want too many trying to block it. I think it could happen in every city in Canada. Last year, McGill produced more startups than any other university in North America.

Key Takeaways

Much of the story that John tells of the early days of the UK research-based ecosystem of the 1980s has clear parallels to present-day Canada. The early development of the UK approach to unlocking the value available in their universities differs significantly from the Bayh-Dole framework that was being developed concurrently in the United States, however, and therein lies the key set of lessons for Canada.

Whereas the United States enacted Bayh-Dole and shortly thereafter the SBIR as government-led push toward private sector involvement in research commercialization while keeping the public and private funding in separate vehicles, it seems the UK, like several other countries, went with a more hand-in-glove model of public-private partnership, a model that to my mind is more readily translated to Canada given that the Canadian emerging technology ecosystem bears more resemblance to that of the UK than it does to that of the US.

The University Challenge Seed Fund (UCSF) was everything that my previous article said that such an early-stage innovation investment fund needs to be: a philanthropic, public-private partnership that focuses on long-term self-sustainability and impact.

Being a public-private partnership allows for risk sharing between stakeholder groups that are otherwise large risk-intolerant. Using a philanthropic model as opposed to a traditional VC-LP for-profit structure to ensure flexibility with respect to the time to impact, avoiding mismatch between long deep tech and emerging tech commercialization timelines as compared to the liquidity requirements often imposed by private investors. Being arm’s-length from the universities ensures sound governance that can otherwise be subject to the whims of a changing administration .

Most importantly, as I laid out in my recent discussions of risk tolerance, it needs to say “yes” a lot: it is much more important in the early stages to ensure that opportunities are not missed than it is to pick winners, and a 50% success rate for investment is extremely high relative to any early-stage investment you can get in Canada.

John takes it one step further, though, in pointing out the value of the due diligence packages that UCeed conducts on early-stage technology. He is absolutely right that such deep due diligence makes no sense whatsoever in a for-profit model and represents a hit to the bottom line of a fund. However, with a philanthropic model where profit is not the primary motivation, activities that support the mission, but which may not be aligned with profits, become fair game. John described UCeed as an economic development engine, and, thanks to the charitable structure of the fund, can focus on a much broader definition of value creation when justifying how money is spent than would be possible in the context of a for-profit VC fund. Metrics like John’s 10-15x multiplier on initial investment within 12 months, while of limited value to a VC’s balance sheet, are clear indicators of progress toward economic development goals. In other words, John’s experience in the UK and with UCeed in Canada both provide clear evidence of the value of the charitable model of support for emerging technology.

What is most remarkable to me about the UK story is what followed the first fund. The fact that this fund took 14 years to reach evergreen status is irrelevant considering the nearly $1 billion (with a B) of private investment capital that is now pointed solely at Oxford university as a direct consequence of that proof of concept. From that seed of £40M GBP invested across several UK universities, so much value and such a vibrant ecosystem of research commercialization sprang up that a single UK university on about the same scale as most Canadian U15 universities now gets substantially more (private, for-profit) investment in commercialization of research than all of Canada’s universities combined. Clearly, this initiative created an enormous amount of private sector value in the form of deal flow that would not have existed without it. These intermediate progress metrics are key to telling the impact story and to starting the flywheel, while the Oxford Innovation Fund (OIF) is proof of the value of doing so, if only we can convince the powers that be to think in generations instead of election cycles.

While John is careful to note that not all universities performed equally well, this is clear proof of the potential value to be created, if we can get the rest of the equation right. On that topic, John’s insights are equally valuable. The key elements, as John points out, are investment, intellectual property, and acceleration. Without the money the rest of it falls apart, but just throwing money at a problem is not enough. To make research commercialization viable, we need to carefully manage the IP arising from publicly funded research, and we need to ensure robust tools for acceleration, which is to say programs that can assist in the early days of business development with entrepreneurial training, mentorship, and networking. While Canada has programs recently stood up to address both IP (see ElevateIP and provincial IP organizations, and my own SAIL initiative) and acceleration (see the i2I and Lab2Market networks), relatively few programs exist that do all three, and those that do are often limited in geographic scope.

A common objection that I hear when discussing anything pan-Canadian in scope is that it will not work because every university is different. Like John, I disagree with this (it is true that every university is superficially different in terms of their IP policy structure and the willingness of the leadership to engage in commercialization activity, but I reject the notion that this precludes initiatives that can work across institutional contexts - it just means a bit of extra work). While my answer to this objection has been SAIL, which is constructed to work with any institutional IP policy and tech transfer setup in Canada, John’s is that we need a pan-Canadian fund that can fill in the gaps and operate flexibly with respect to institutional approaches to tech transfer if we are to achieve the same kind of impact as was achieved in the UK; a fund that, given the risks and timelines involved, should be closely modelled after UCeed as a public-private, philanthropic, evergreen fund that can act in service of a greater mission than just profit. SAIL provides a foundation on which that can be built.

A structure that requires a generation to achieve evergreen status, is a difficult sell in an ecosystem where leadership at every level changes every few years.

The path forward involves a multidimensional approach, part of which lies in culture. John alludes to this idea in a number of places, most importantly when discussing the idea of faculty identifying with the missions of specific sub-funds of UCeed, and of using faculty engagement as a performance metric for Innovate Calgary. This is in line with the lessons learned through the Promotion & Tenure, Innovation & Entrepreneurship (PTIE) initiative about which I have previously written, wherein a grassroots culture shift, seeded deliberately at 75+ American universities, led to dramatic changes with respect to the impact of research beyond the lab. A top-down mandate to commercialize is important, but if we are to create something with the resilience to survive long enough to achieve the kind of impact that turned the UCSF into the OIF, it needs to be predicated on institutional culture, a requirement for which PTIE provides the blueprint.

As John points out, the cost of moving forward is small. The initial cost of the University Challenge Seed Fund, converted to CAD at the 1999 exchange rate and adjusted for inflation since then, amounts to about $175M CAD. Compared to the $8.5B that our unmet NATO commitment says should be spent on emerging technology development (20% of our 2% commitment, which is soon going to be boosted to 5%), or to the amount currently being spent under SR&ED (about $4B), or the amount that France is putting into this problem through Les Deep Tech (€3B EUR over 10 years), this is nothing.

I will end by highlighting one key element of John’s commentary: right now, we need action, not more studies. We understand the problems. Tried and tested solutions exist. We don’t need more papers written about it, we don’t need more reports. We need leadership from both the public and private sectors willing to embrace risk. UCeed is doing excellent work, but without an equivalent with national scope, we will continue to leave potential value on the table. John makes a strong case for the value of redirecting even a small fraction of the money that Canada is currently spending on blunt instruments like SR&ED toward more focused initiatives to support emerging technology arising from publicly funded research.

Many thanks to John, both for taking the time to discuss this with me and for his leadership in Canada’s emerging technology and innovation space.

One of the most powerful forces for systemic change arises when a top-down mandate to reach some target outcome meets a bottom-up movement of people with domain expertise willing to do the work, squeezing out the sticky middle that usually exists in opposition to such change. Examples abound, but probably the most salient one I’ve discussed to date newsletter arose from the meeting of Bayh-Dole’s mandate to commercialize public funded research with PTIE initiative that has transformed more than 75 American research institutions into innovation engines in recent years.

The last couple of weeks present such an opportunity for Canada. PM Mark Carney’s mandate letter to his ministers arrives close on the heels of a number of grassroots initiatives to change how Canada approaches innovation.

The 7 priorities in the PM’s letter lay out high level target outcomes without getting into the weeds of how. This is as it should be: in practice, a top-down mandate usually works best when it prescribes target outcomes and the reasoning behind them (the what and the why), while implementation details (the how) are left to the people on the ground who have the domain expertise to deliver. Below, I comment on each of the 7 priorities highlighted in the mandate, pulling from my work as well as that of CCI and others to provide some perspectives on what I hope to see on the implementation side.

The stars appear to be aligning toward real, systemic change, as all the conditions exist to make this possible - but it remains to be seen if this government can follow through.

Reading between the lines

PM Carney’s letter is short and to the point, as is most of his communication to date.

In the first few densely-packed paragraphs, PM Carney positions Canada within a quickly changing international trading system, calling on the ministers to redefine our commercial and security relationships, become an energy superpower, address the housing crisis, and secure our borders. He dedicates two entire sentences to AI adoption, an enormous emphasis given the scope of what is passed over as a set of items in the list above.

Before we get into the content, it’s worth flagging what is by far my favorite part of the letter: all the ministers were all sent the same letter.

This has been a somewhat polarizing choice, from the discussion I have seen so far, with the main criticism being that by not clearly delineating responsibility for these missions, the PM risks conflicts over jurisdiction. Far from being a problem, this is both intentional, and brilliant.

A system as large and complex as the Canadian government does not exist in siloes, no matter what the org chart says. Any decision made at the ministerial level has ramifications beyond just that portfolio.

The message that underlies the uniform mandate letter is loud and clear: all of these missions involve all ministries. As the innovation community has been saying since at least 2011, if not before: none of what needs to happen to change Canada’s trajectory will be possible without whole of government leadership. It is my sincere hope that this marks the beginning of the dismantling of ministerial siloes that has plagued Canadian policy development in recent years.

A (set of) Generational Challenge(s)

Priority 1: Establishing a new economic and security relationship with the United States and strengthening our collaboration with reliable trading partners and allies around the world.

This first point is entirely as expected, and in line with what his government has been doing since he took over from Trudeau. Obviously, the scope of what is involved in this sentence is vast and I won’t attempt to fully unpack it, focusing instead on the role played by intellectual property in delivering on this priority.

Intellectual property is what exists in the intersection of economic and national security with innovation policy. I have written previously about the importance of a thoughtful approach to foreign direct investment, an issue which takes on entirely new dimensions in light of recent American collective insanity policy shifts. CCI gets into it as well, calling on the new government to “protect Canadian ideas and our capacity to innovate by taking a sovereign, security-informed approach to IP, data, and other intangible assets to maximize freedom to operate.” as well as to “Better protect Canada through economic security by bringing the Investment Canada Act into the 21st century.”

This is also a mission to which the work of the SAIL team is intended to directly contribute by making it easier for Canadian startups to access valuable intellectual property arising from publicly funded research. The Simple Agreement for Innovation Licensing (SAIL) and the rapidly evolving set of projects building on the framework seeks to create conditions in Canada that promote domestic value creation and retention of valuable intellectual property by making it possible and desirable to build in Canada. All of this will contribute directly to security issues, but under a siloed mandate would fall under innovation policy rather than security policy. With a mission that spans ministries, the scope and potential impact are greatly broadened.

Priority 2: Building one Canadian economy by removing barriers to interprovincial trade and identifying and expediting nation-building projects that will connect and transform our country.

Again, entirely consistent with what was already happening. The CCI report similarly calls for “tackling internal trade barriers and creating a single market for public procurement from coast to coast to coast”.

While this priority has been heavily focused on barriers to interprovincial trade, there are two other key elements that fall under the heading of “nation building” that I would like to highlight here.

The first relates to emerging technologies that require multi-level engagement and scale to demonstrate utility. Climate tech (specifically carbon capture) is a perfect example: while most truly innovative carbon capture tech is being pioneered by small startups, it is only by reaching massive scale that they can be relevant to the problem they seek to solve. This is not trivial: Park et al. do a very good job discussing the challenges to scaling climate and cleantech, flagging a need to coordinate across multiple stakeholder groups at the municipal, provincial, and federal government level and in the private sector. These represent enormous internal barriers to adoption of innovative technologies. In climate tech, the private sector is already leading on this front, with Deep Sky providing a scaleup environment for an enormous variety of carbon capture technologies that bears some thematic resemblance to the foundational work that established Canada’s quantum valley in the early 2000s.

The second focus area must be on removing the competitive nature of provincial support for innovation. A good example of this lies in the existence of early stage investment tax credits in some provinces but not others (30% BC, 45% SK and MB). Every province should have these, and they should be the same everywhere, or at least offset at the federal level to bring them effectively in line. We should not be competing internally to attract talent from other provinces through zero-sum internal competitions. We should be focused on competing at a national level with the rest of the world to bring talent to Canada.

Priority 3: Bringing down costs for Canadians and helping them to get ahead.

I’m not sure this one deserved its own priority line, and the vagueness of this priority contrasts oddly with the focused and specific tone of the rest of the letter.

While it does not hurt to make this explicit, all else being equal this will be a natural byproduct if the rest of the priorities are delivered.

Priority 4: Making housing more affordable by unleashing the power of public-private cooperation, catalysing a modern housing industry, and creating new careers in the skilled trades.

Housing is directly connected to innovation issues. I’ve written previously on the connections between the lack of private sector investment and the housing crisis: investment allocation across asset classes being zero-sum, any incentive to invest in real estate is a disincentive to invest in innovation. recently put out an excellent piece discussing the disincentives to innovation created by high profit margins in stagnant and uncompetitive industries, and housing certainly qualifies as sector that Canadian policy has allowed to turn a complete lack of productivity into what is effectively a regressive generational wealth transfer.

More explicitly, one of many reasons that Canadian private sector capital is risk-averse is because they can make so much more money investing in safe real estate or by continuing the uncompetitive status quo. Anything that makes housing less attractive as an investment will directly incentivize investment in other sectors while driving down housing costs.

The mention of public-private cooperation is also key, and I am somewhat disappointed to see it buried in a single priority. I have long (relative to the existence of this newsletter, that is) advocated for the importance of public sector leadership as catalyst to innovation, and public-private partnerships are central to proven risk-sharing models that work elsewhere. While it will certainly be an important part of addressing the housing crisis, I would have liked to see this element of the mandate elevated to a more general position that spans all of them.

Priority 5: Protecting Canadian sovereignty and keeping Canadians safe by strengthening the Canadian Armed Forces, securing our borders, and reinforcing law enforcement.

In recent history, there has been no greater drive of innovation and economic development than the threat of war, and if we must deal with a more volatile geopolitical landscape, let’s at least not waste the opportunity to build. Defense procurement and the building of a Canadian defense industrial base is an enormous opportunity for innovators and those seeking to commercialize emerging technologies.

Canada’s first-in-class quantum technologies, our untapped ability to project geopolitical power through food production, our foundational work in AI development, our CANDU reactor technology, and the host of dual-use technologies that are emerging from Canadian research institutions could all contribute to a renewed focus on domestic defense procurement.

Priority 6: Attracting the best talent in the world to help build our economy, while returning our overall immigration rates to sustainable levels.

I wrote recently about the immediate opportunity before us to poach American scientific talent. As they seek to flee a regime hell-bent on imposing ideology on science, Canada stands to benefit enormously.

Almost.

As I noted in my previous article, we do not actually have in place any way to turn the output of these scientists into economic and social benefit, and without that, there is not much point. Dan Breznitz went somewhat further in a series of op-eds to the Globe and Mail, he makes the point that Canada’s total factor productivity, the nonlinear increase in productivity that arises above and beyond the linear contribution of additional capital and labour, is effectively nil, and was actually somehow negative from 2000-2008. Let that sink in: we have been so inept at effectively employing the people we bring into Canada that we have actively hindered their contribution to productivity.

If priority 6 is to serve the interests of priority 3 and 7, the outcomes laid out here are far from being isolated to immigration. They touch on housing, on innovation, on economic policy, and on security. Good thing all the ministers got the same memo.

Priority 7: Spending less on government operations so that Canadians can invest more in the people and businesses that will build the strongest economy in the G7.

Having laid out all the ambitious priorities to this point, this final priority lands very effectively.

While the word “innovate” appears exactly zero times in the letter, this last point makes it clear that this is a mandate to innovate (well done, Laurent, on the prophetic naming of CCI’s report). Not only are the ministers being asked to deliver solutions to multiple, intersecting generational challenges, they are being asked to cut costs while doing it. In other words, this set of priorities demands outside the box thinking. It will not be possible to deliver on any of this while reducing costs, without fundamentally changing how things get done.

Delivering on this will require acting on a large number of low-hanging fruit: building efficiency into public service offerings. The EU’s Only Once Principle and related implementation would be an excellent example to follow that also ties directly into Priority 2. AI adoption will also likely be top of mind given the top billing it got in the mandate letter.

Wrapping Up

The ministers have their mandate, but to deliver it, they will need support from people on the ground with domain expertise to address implementation details. There is no shortage of bottom-up support on the execution side. The CCI report, A Mandate to Innovate, lays out clear action items for ministers, offering a highly dense compendium of specific action items with which I mostly agree (to be picked apart in a future post). Upstream from CCI’s focus, the SAIL initiative serves as a vital grassroots movement to fill the input end of the economic development funnel and drive value creation from publicly funded IP. An enormous number of people and organizations have weighed in on topics like SR&ED reform, patent boxes, the capstone agency, the Canadian innovation corporation, and more (See QIC, CCI, FORPIQ, and my own submissions here and here for a (very) non-exhaustive sampling). The challenges are, for the most part, well understood and solvable across the entire innovation pipeline. Translation into policy is the bottleneck.

With these mandates set in motion and momentum steadily building from the ground up, let's hope the government can navigate the coming challenges by listening to the people on the ground, and avoid the risk of continuing Canada’s aimless drift through increasingly stormy geopolitical waters.

This article is a companion piece to my recent article “Embracing Risk”. While not strictly necessary, if you have not already, I encourage you to read that one first to place some of the ideas presented here in context.

You have probably heard, at some point, that Canadian startups on average raise less money and return smaller multiples for investors than American startups. The Logic gets into it, as do many others. The natural but probably wrong conclusion is that Canadian startups are less capable, in some sense, than their American counterparts. The problem with the simplistic conclusion relates to sample size.

Once upon a time, I got into a friendly argument with a colleague over chess rating statistics while drinking the truly awful beer that fuels grad school. Their assertion was that, given that men make up the vast majority of top-ranked chess players globally, men must be better than women at chess. I disagreed, on the grounds that there are far fewer women who play chess professionally than men, and that one would expect that underrepresentation to be reflected at all levels of play.

Being the nerd that I am, I had to prove my assertion with data to satisfy my own curiosity. You can see the ensuing Stack Exchange thread here. I will not reproduce the analysis, but in a nutshell, the data bore out my argument. Men outnumber women in rated chess play by a ratio of about 23:1. When you sample a distribution, the number of outliers you will draw increases with the number of samples you draw, and so you expect that the top spots to be dominated by samples from the larger population.

When comparing outlier performance between two populations, it is not enough to simply compare directly and conclude that the population that has more and larger outliers is represented by a different underlying distribution with a larger average, unless you are certain that the sample sizes are the same. Even two samples drawn from the same underlying population can exhibit significantly different performances when sample sizes differ.

[[As an aside, the quick and dirty analysis I did at the time suggests that top-10-ranked women actually tend to overperform expectations once sample size effects were accounted for. To put this in more concrete terms, all else being equal (which is to say, if the imbalance between men and women remains the same as it currently is), we expect a Magnus Carlsen-level male chess player to be born every 30 years or so. In contrast, given the level of underrepresentation of woman among grandmaster chess players, Judit Polgár is a once-in-a-millennium chess mind.]]

Under-sampling creates an illusion of relative underperformance

The demonstration of the impact of chess ratings suggests that we should be very careful with comparison of population outliers, which are what matters when considering startups. Given that there are fewer Canadian startups than there are American ones, and given further that startup performance is dominated by outliers, we are likely making the same mistake as many do when considering chess performance. Apparent Canadian startup underperformance relative to American startups may simply be a sample size effect.

In other words, it can be true both that American startups and Canadian startups are equally performant (in the sense of being drawn from the same underlying value distribution), and that a Canadian startup will probably never be top-ranked by valuation at any given time when considering both ecosystems together. The resolution of the apparent contradiction lies in the fact that there are just more American startups than there are Canadian ones.

No comparison that I have ever seen of the two ecosystems has attempted to take this into account. Various explanations of this have floated around for a while, including some particularly silly ideas about culture, but I have not yet come across any attempt to rule out the most obvious issue: that the sample sizes are vastly different. While I do not have the data to find out if this is able to explain the entirety of the problem, it is certainly part of it, and, as I argue in the rest of this post, it is probably a core element of the cause of issues that compound the problem.

Perceived under-performance is a self-fulfilling prophecy

It gets worse, because perception of underperformance is a self-fulfilling prophecy. Investors considering allocating their assets care only about measurable performance, meaning that an investor with the option to invest on either side of the border and no preference, all else being equal, will invest preferentially in American companies, increasing the availability of capital and increasing the chances of success in that ecosystem - effects which will drive real differences in the underlying distributions. In other words, under-sampling creates a negative feedback loop in which the expectation of performance differences makes them real. This prophecy is played out in several ways, including difficulty raising money in Canada, as well as a tendency for highly successful Canadian founders to have companies that are based in the US.

Under-sampling power laws causes real underperformance

When considering return on investment for startups there is another, much less intuitive complication not present in the case of chess ratings that needs to be considered, which arises from the asymmetric nature of the power law. While the distribution of chess ratings is symmetric about the mean, the power law that dictates startup value is about as asymmetric as it is possible for a distribution to be. This matters, because when we draw a sample population from a symmetric distribution, the distribution of the population average is itself symmetric, and its median and mean therefore independent of sample size.

In an asymmetric distribution like the power law, this is not the case. The distribution of the population average remains asymmetric up to quite large sample sizes, and smaller populations actually have a distribution of population averages that is skewed toward smaller numbers. The smaller your sample size, the worse the average performance of that population, and vice versa. (For the stats nerds out there: the central limit theorem guarantees that for a sufficiently large sample size, the distribution of the average will converge to a symmetric Gaussian distribution, but different underlying distributions dictate very different values of “sufficiently”, which matters here.)

Restating this in investment terms: if you are investing in power-law distributed assets like startups, up to a point, you can increase the median performance of your portfolio simply by making it larger. The converse is also true: by making fewer investments, you are actually driving a real, absolute reduction in expected outcomes. This is not just an illusion of relative underperformance caused by under-sampling, it is real underperformance driven entirely by sample size.

The following plot shows the impact clearly. Using the same distribution parameters as presented in the Moonfire article for the multiplier on ROI, I draw 10,000 examples of populations of the size given on the x axis, and for each population size I plot the boxplot of the distribution of the sample average of these “portfolios” (you can interpret the y axis as the average multiplier returned by each company in a portfolio of N companies, with N given on the x axis, with the boxplots showing the quartiles of the resulting distribution portfolio averages).

A value of 1 on the left y axis is the point at which an investment portfolio breaks even, on average, but the absolute numbers are not all that important (they are drawn from a distribution with realistic but arbitrary parameters that reflect qualitatively the underlying distribution, not necessarily quantitatively, and we are obviously ignore the impact of time in this simplistic analysis). What is important is that the median of the distribution of portfolio averages (orange line) is an increasing function of portfolio size until we reach portfolio sizes that are orders of magnitude larger than typical VC portfolios.

We tend to implicitly assume that averages are independent of population size, while variability goes down as it gets larger, but this is only true for symmetric distributions (or for very large sample sizes). For power laws and other heavy-tail distributions with smaller sample sizes, or any situation in which it is the outliers that we care about, both the sample average and sample standard deviations are strongly sample-size dependent until the sample size is very large.

The secondary y axis on the right shows the fraction of population samples that have at least one portfolio company that returns a multiple that exceeds 100, which we use as a proxy for the fraction of portfolios that hit at least one homerun. You can see that once for small sample sizes you are all but guaranteed to miss, but as this probability goes up, the median performance approaches saturation, which is just another way of seeing that all that matters when making investments is not missing the homeruns.

For this specific example distribution, a Blind Squirrel investor needs about 300 investments to be made before the sample size effect goes away and average portfolio company performance saturates. This is the core of why the Moonfire model found that best performance occurred when deploying all their capital in the first round - by considering portfolio sizes below this threshold, they are operating in a regime where extra portfolio performance could be extracted simply by making more investments.

With power laws, under-sampling causes real underperformance.

How does this get reconciled with traditional VC?

The obvious question, given the above, is “how can VCs still operate with portfolios of 20-50 companies?”. For a given niche, and at a later stage, it is actually possible to pick winners to some extent. What this means in mathematical terms is that in some narrow circumstances, it is possible to change the underlying distribution to a power law where average performance saturates for smaller portfolio sizes (mostly this is about reducing the exponent in the power law). B2B SaaS and biotech, the two sectors where VC has proven effective, are these niches. Later stages of investment across a broader set of companies (series A and beyond, for example) also have distributions with reduced exponents. In these cases, VC can be successful because it is possible, to some degree, to pick winners and bias the pool of possible investments toward a more forgiving power law.

That being said, I suspect that in most cases, VC performance would similarly be augmented through larger portfolio sizes, which may be one of the drivers behind the trend toward larger and larger VC funds generally, though compensation structure also plays a role there since the 2 and 20 rule increasing frontloads reward as fund size goes up. The limitation here is due diligence times - changing the exponent that drives the power law is a lot of work.

This generalizes poorly to making early investments in emerging technologies and deep tech, though, and this is the reason that traditional VC does poorly here relative to B2B SaaS. Picking winners (by which I mean reducing the power law exponent) is much more difficult in the earliest stages and in sectors for which the metrics used by traditional VCs to perform valuation break down. Here we are at the mercy of the underlying power law, but we can use our statistical understanding of the impact of sampling to correct for challenges with picking winners.

Emerging Technologies

Where emerging technologies are concerned, if we acknowledge that there is not much we can do to predict technology portfolio value early, the only winning move is to take the Blind Squirrel approach and invest broadly. This is the core of why I am a strong advocate for risk-tolerant, philanthropic investment approaches that partner public and private sector capital toward a common goal. These approaches can tolerate long timelines, can consider the value of positive spillovers that are missed by traditional VC, and, most importantly, can place lots of bets. Les Deep Tech in France are taking this approach, with am ambitious goal of creating 5000 deep tech companies over the next 10 years. I do not need to know the details of the value distribution from which they are drawing to know that they are operating comfortably in the regime of maximum average performance. The SBIR in the US similar supports thousands of emerging technology companies every year, with excellent aggregate results.

For policy makers and those seeking to support emerging technologies, the important point is to recognize that underperformance of Canadian deep tech (either relative to the Americans or in absolute terms) is the result of a self-fulfilling prophecy that reinforces the systemic risk-aversion that leads to that underperformance in the first place. It is at least partially, and possibly entirely, because of risk-aversion-driven under-sampling that Canadian startups underperform. Breaking this cycle requires an active choice to take more risks, and the math tells us that we will be rewarded for doing so.

In other words, our innovation policy frameworks have cause and effect precisely backwards. Risk-intolerance is a direct cause of real underperformance, not the other way around. Canadian deep tech and emerging technology startup underperformance can be remedied simply by embracing risk and investing broadly, and staying the course long enough to see the impact.

Today I give you a break from my usual innovation policy ranting discussion to announce the release of SAIL version 3, as well as a major facelift for www.howtosail.ca. This website will serve as a central hub for all things SAIL-related going forward, including copies of the latest documentation, a detailed archive and changelog, and a list of partner organizations that are involved in building and using the SAIL framework.

If you are interested in getting involved with the SAIL initiative as a supporter, a contributor, or a user, or are simply interested in staying up to date, please get in touch. You can also sign up on the website to receive updates and news relating to SAIL going forward, or follow along on LinkedIn.

SAIL v3 changelog

The major changes between v2 and v3 are concentrated around the financial model. What we heard clearly from those who provided feedback on v2 was that the components of the convertible debt included in SAIL required accounting overheads that added unnecessary complexity. To address this, we have separated the components of convertible debt into two buckets, one of which requires careful accounting and corresponds to accounting that universities were already doing anyway, and one that does not. In this way, SAIL v3 should be no more administratively complex than any existing licensing framework.

Alongside this version, we have also significantly expanded the SAIL guidance document. The guide, also available on the website, provides a detailed rationale, grounded in the axioms of tech transfer and discussions with stakeholders from across the country, for every decision made in SAIL. The guide walks users through the process of a SAIL negotiation, section by section, clearly delineating responsibilities and elaborating on the spirit and intention behind the various clauses in SAIL. We note that the guide is not yet in plain language, which will be corrected in the next few weeks.

The other significant change relates to the axioms themselves. A few axioms have had wording tweaked to clarify intent, with the most significant changes being made to axiom 5, which was, in hindsight, not very clearly articulated in SAIL v2. This axiom relates to how institutions should seek to benefit from supporting commercialization activity. SAIL’s position is that the amount of convertible debt and eventual equity being tied to the marginal cost of commercialization above and beyond the costs involved in the research process itself. Given that the institution has already been paid through other means (mainly taxpayer-funded grants) for the work involved in undertaking the research and disseminating the results, only costs that are not directly aligned with their mandate, such as patenting, market studies, IP management support, and other commercialization-specific costs, should be reflected in institutional equity stakes. Just as with any angel investor that is rewarded with equity that reflects the cost of their investment, SAIL provides a framework through which universities can simply account for their costs and see reward that is proportionate to the support they provide.

The guidance document dives deep into the rationale for this, also providing a framework for thinking about situations in which it is reasonable to execute a license agreement that may not be compliant with all of the axioms.

Definitions and language have been edited for simplicity and clarity throughout. While it is not yet at the standard of “plain language”, it is getting very close. You can expect a minor release to take it the rest of the way in the near future.

The last piece of the puzzle is funding, and providing research institutions not just the tools (SAIL) to deliver impact beyond the lab, but also the resources. Stay tuned for developments on this front.

Acknowledgements

Many thanks to University of Ottawa Innovation Support Services, for hosting the roundtable discussion that informed the latest update to the framework using funding from Intellectual Property Ontario, and to co-hosts MVIP™ Solutions, Inc. and FORPIQ, to Robin Ford for her work simplifying the language in the agreement enroute to a true plain language contract, and to all those who participated and shared their insights.

Many of my recent posts here have been themed around the idea of risk tolerance. This post attempts to put some mathematical weight behind the assertion that embracing risk is a prerequisite for value creation through innovation, and makes the case that public sector leadership is required if Canada is to ever secure socioeconomic benefit from publicly funded research.

If you are involved at all in the startup or VC space you have no doubt heard all about power laws. In a the context of emerging technologies, a power law distribution of technology value means that the vast majority of the value arises from a small minority of technology portfolios. Unfortunately, it is very difficult to tell the difference upfront, and the earlier you seek to get involved, the harder it gets.

In this post, I elaborate on the implications of this value distribution for public-sector risk tolerance for innovation funding, demonstrating how Canada’s strategy of attempting to pick winners results in exactly the opposite. I discuss the origins of systemic risk-intolerance that has prevented public embrace of the power law as a means to benefit from Canadian emerging technologies, and highlight ways other ecosystems are thinking about creating value from emerging technologies.

Together, it all suggests that a shift in the way we think about the value and role of public sector spending is a necessary precursor to addressing Canada’s misnamed Productivity Paradox.

The VC model has been around for a while

Venture capitalists (VCs) learned how to navigate power law dynamics a long time ago, and the basic structure of the approach persists to this day. The phrase has its origins in early 19th century whaling expeditions. Agents (analogous to modern VCs) would raise money from corporations and high net worth individuals (LPs) to fund ship’s captains (founders) to go whale hunting (ventures). Even the distribution of payout (the 2 and 20 rule) arose during this time.

As with modern investments in emerging technologies, the vast majority of whaling expeditions ended in failure, while the top 2% returned so much value that they more than paid for the rest. Without any real way to predict which ones would return profits when they set sail, the only winning move was to fund many expeditions, enough to ensure (statistically speaking) that at least one of the investments would be among the 2%.

The process has changed since then, but the underlying model is still effectively the same.

Power law math

Simulation modelling of power law dynamics has been done well by a number of people: Matt Lerner over at Medium developed a very simple yet enlightening toy model that simulates random investment with power-law distributed returns, while Mike Arpaia from Moonfire developed a more sophisticated model as a means to test variations on early-stage investment theses. Rather than reproduce their analysis, I will simply summarize their findings here. I encourage you to skim these articles before reading further if you are not already familiar with power law dynamics.

In a nutshell, Matt’s Blind Squirrel model does a Monte Carlo simulation of portfolio values over random investments made in prospects with a power law distribution of returns. Mike Arpaia builds a more sophisticated framework that allows comparison of different fund structures and allocations, with the obvious goal of optimizing payout from early-stage investments, but the core of the model is the same: an underlying distribution of investment outcomes is sampled according to a set of rules that collectively define an investment thesis, and the hypothetical returns are compared.

In both cases, it quickly becomes clear that even if all you do is throw darts you’re all but guaranteed a positive return as long as you build a sufficiently large portfolio to have a reasonable certainty of hitting a few top performers. The Blind Squirrel approach achieves this simply by making a lot of bets and letting the law of averages sort it out, while the Moonshot model advocates for deploying most of your investments in the first stage, with minimal reserve for follow on rounds, for the simple reason that this allows more bets to be made.

While the Blind Squirrel approach underperforms average VC performance, it is still net positive. The implication is that you do not need to be able to pick winners to have reasonable certainty of return when playing a power law, as long as you make enough bets to cut through the noise.

The numbers-game nature of early-stage investment comes across most clearly when considering the impact of the top performer in a given portfolio. In the Blind Squirrel model, if you remove the top company from each portfolio, the returns drop dramatically, with the top company providing as much as 50% of the profit for small portfolios. This dependence on the top performer gets less and less important as the number of investments made by your fund increases.

In other words, if you are running a small fund and you pass on the one big opportunity, it’s the difference between wild success and complete failure. As a Blind Squirrel, the only winning move is to invest in a much larger pot of companies, and let statistics take care of picking winners. The Moonfire model arrives at broadly the same conclusion.

This is not a particularly controversial or surprising conclusion. We are basically just restating the hypothesis that portfolio diversification is important. Anyone investing in a broad market ETF is effectively a Blind Squirrel investor, albeit using a different asset class.

When considering emerging technologies, the power law gets even more extreme, the timelines get extended, and it gets even harder to pick winners given that there is an entirely new class of risk involved. Not only do we need a team that can deliver and a willing market to buy, in many cases we also need the laws of physics to cooperate with the research efforts. With the added filter, it takes a lot more sampling to cut through the noise. If the top fraction f of companies or IP portfolios are responsible for most of the value creation, a Blind Squirrel needs to build a portfolio of 5/f companies to ensure a 99% chance of hitting at least one big winner (that’s 250 investments, assuming 2% of companies are the profit-creating outliers). The earlier an investment is made, the smaller f becomes.

The conclusion is simple: successful incubation of emerging technologies requires a support structure that is both willing and able to make a very large number of bets, understanding and accepting that the majority of them will fail. All that matters from a profit perspective is aggregate performance, which is dictated almost entirely by a small minority of investments. In many ways optimal investment strategy resembles a midwit meme, with VC as the midwit picking winners while both very early and very late stage investors are best served by just buying (almost) everything.

Drivers of public sector risk tolerance

There is a fascinating piece of research that gets into the dynamics of the unavoidable tradeoff between false negatives and false positives in public spending at a cultural level, which you can read here. The authors explore cultural differences in consideration of “second-best fairness”, the idea that there is “a trade-off between giving some individuals more than they deserve, false positives, and others less than they deserve, false negatives.”

While they focus on welfare payments, the idea generalizes well to anything that involves spending of public funds. If we optimize our systems to reduce the possibility of making a payment in error, we will end up refusing payments in error at least some of the time (false negative), which can have severe consequences for the individuals involved, or be the reason a startup moves south. On the other hand, if we optimize to ensure that payments are made to those who need them, we will make some payments in error, whether honest mistake or the result of fraud (false positive). It is not possible to build a system that completely eliminates both of these, and the balance selected is an active policy choice. ( did a great Statecraft interview recently that touches on these issues).

Canadian tax code is a good example of a system optimized to avoid false negatives. CRA assumes your return to be accurate, a few basic checks aside. There is some degree of errors and fraud occurring at any given time as a result, but it is only worth correcting if the amount recovered by doing so exceeds the cost recovery, which leaves some false positives as acceptable outcomes.

Where innovation policy is concerned, Canada optimizes heavily to avoid false positives. If you have engaged with almost any public funding for innovation you find that these policy frameworks require that a success narrative be told for every project funded, irrespective of their aggregate impact. did a good job summarizing the core of the issue in this recent post, in which he suggests that many policy decisions are made not to ensure maximal benefit, but to avoid blame for failures. No politician or public servant wants their name attached to a project that will not bear fruit in their term, since it becomes an easy target for opponents.

Picking winners

When the goal is to avoid the possibility of blame for failure at the level of individual investments, we lose sight of the broader context in which that investment occurs, a problem that is exacerbated by the lack of unifying mission that characterizes Canadian innovation policy space generally. Our innovation support programs are incentivized to avoid individual failures by only supporting low-risk projects, which in turn ensures low reward.

We try to pick winners, and we are not very good at it.

Many programs have minimum revenues or headcounts to be eligible, for example, the implicit assumption being that revenues (or jobs created, or T4s issued, or any numbers of other short-term measures of economic impact) represent de-risking of an idea. This is a problematic assumption when considering emerging technologies, which are often years of research away from any revenue and are often best advanced by small, agile teams. Where emerging technologies are concerned, the signals that our innovation support systems use to pick winners are uncorrelated to long-term impact until far too late in the game.

Power law math tells us that Blind Squirrels cannot afford risk aversion. The power law that dictates the value of emerging technology portfolios, combined with Canadian systemic public risk aversion, means that much of Canadian innovation policy is constructed to miss out on value creation arising from research. Canada’s Productivity Paradox is anything but paradoxical: it falls naturally out of even the simplest toy model we can conceive.

The models and discussion above make a key assumption: that our investment strategy does not alter the underlying distribution of investment returns. In reality, any program that invested in literally everything would immediately be subject to fraud that would drive the value of the underlying distribution to zero. Some degree of selectivity is required, but the math suggests that, at the earliest stage of technology development, the selection should not be much more stringent then filtering out obvious fraud, making sure the founder is both credible and coachable, and ensuring that the amount requested is aligned with the actual need.

Value is not synonymous with profit

While the track record shows that on average VCs do outperform the Blind Squirrel investor, the effort required limits a typical fund to 20-50 investments, and having profit as the primary driver limits the timespan over which those investments can be held. Both of these limitations make traditional VC poorly suited to supporting emerging technologies, where timescales are extended and picking winners is all but impossible. It’s also a vicious cycle: the need to pick winners increases due diligence requirements, which in turn further limits the number of investments that can be made, further increasing the required success rate. It works for B2B SaaS and biotech for which playbooks have been developed to assist in the process, but it translates poorly to other sectors.

There are two sources of investment that have the ability to truly play the numbers game while being tolerant to long development timelines: venture philanthropy, and the public sector.

To make early investment strategies compatible with securing socioeconomic impact from emerging technologies, we need to expand our definition of value creation to recognize that value creation is not synonymous with profit. If instead of pure profit-seeking we expand our definition of value creation to include economic development, security, and independence, retention of talent and IP, education of entrepreneurs, and progress toward ambitious societal goals like climate change targets, the Blind Squirrel approach becomes much more attractive. While such positive spillovers are of no use to a for-profit VC’s balance sheet, they are of clear value to the Canadian public and to mission-driven investors.

Unlike traditional VC that is limited to just a few investments over a fixed timespan and all that matters is (some metric of) profit, public or philanthropic investment in emerging technologies can afford to place many bets with patient capital, all but guaranteeing that opportunities are not missed. In a model that seeks economic development and social impact over profit it suffices to be evergreen, which is an easy target to hit given a sufficiently large portfolio.

Even companies that a for-profit VC firm would write off as failed bets have value in this model. A failed entrepreneur is now someone with invaluable entrepreneurial experience who is better equipped to navigate the process on round two and is incentivized to stay in Canada to try again, knowing they will be supported. Companies that return 2-5X in the long run, while complete failures on a VCs balance sheet, are all contributors to a strong and resilient domestic economy based on SMEs. A breakout success can easily be the basis for a cascade of positive spillovers as entrepreneurs become mentors and investors in the next generation (the story of the transformative impact of the Skype acquisition on the Estonian ecosystem is a great example of this).

A new approach to venture support

There is a desperate need in Canada to embrace the idea that a high failure rate in supporting emerging technologies is perfectly acceptable, so long as the aggregate, long-term impact of the whole portfolio is net positive, and to expand the definition of “impact” to include positive spillovers beyond direct return on investment.

Other ecosystems are way ahead of us.

DARPA has an 85-90% failure rate over its lifetime and the SBIR (“America’s seed fund”) has failure rates that exceed 90% in some cases. Far from being embarrassments, these programs have existed for decades and are widely recognized as cornerstones of American technological dominance. Early-stage risk-taking is a common feature of many other ecosystems that innovate effectively, as well.

A more directly comparable example to Canada is France, where Les Deep Tech has set an ambitious goal of “500 startups created every year, 10 deeptech unicorns, 50 new industrial sites per annum” by 2030, and has $3B in public funding committed to achieving this. Some quick math shows that they are comfortable with targeting a 2% success rate (10 unicorns out of 500 startups per year). They are taking the Blind Squirrel approach of making thousands of small, low-overhead bets, understanding that positive spillovers will ensure net positive value creation.

This is the kind of commitment Canada needs to move the needle in deep tech and emerging technologies, but to make this possible, we first need to embrace risk and recognize that aggregate value creation is more important than individual project success or profit. To do so the Canadian public sector must accept that a high failure rate is an essential and intentional feature of an effective innovation ecosystem and adopt a mission-driven, whole-of-government approach to enacting this in its innovation strategy.

My last article was about learning from our mistakes. That publication was, coincidentally, followed literally the next day by the announcement that two BDC funds were being shut down: IP-back Financing, and the Deep Tech fund.

The writing was on the wall for the Deep Tech fund for some time. When partners leave funds, it is never a good sign. However, the announcement and lack of apparent successors is problematic.

As reported by The Logic, the cancellation was accompanied by a public statement from BDC that reads as follows:

“Overall, this is a normal part of businesses reorganizing themselves to better serve their clients, which included some positions becoming redundant. Like our clients, BDC is becoming more productive, thereby allowing us to serve more clients while reducing our costs.”

Translated for those of us who may not speak corporate, this means that there will not be any public failure analysis.

This is a problem, for all the reasons I wrote about previously, but also because apparently BDC has not deprioritized deep tech and is “working on the thesis and mandate” for a second deep-tech fund. If that is the case, then we need to understand what went wrong with the first version if we are to have any hope for a different outcome in round two.

In this article, I will start the kind of failure analysis I would like to see done for the Deep Tech fund, drawing on the information available publicly and my own experience with deep tech commercialization. An analogous analysis of IP-backed financing will be deferred to a later article.

Investing in Deep Tech

The Deep Tech fund was established only a few years ago as a $200M fund aiming to invest $5M cheques in 15-20 deep tech companies. Originally meant to be a 12-year fund with the possibility of a 4 year extension, closure early represents a major blow to Canada’s commitment to deep tech (or perhaps, an affirmation of its complete lack of commitment to deep tech).

Reading between the lines of the public statements about the closure, it seems the closure was less a failure of the Deep Tech fund specifically as it was a reactionary and counterproductive response to BDC’s severe underperformance as an organization in aggregate, with losses totaling $1B reported in the two years leading up to the announcement. Even so, it is nevertheless worth reflecting on how it could have been done better, in case BDC follows through with the promise of a second iteration.

Timelines

Traditional VC operates on a 10-year cycle, with liquidity requirements starting 5-7 after a fund begins deployment. Deep tech, gated as it is by long development times and research, is often simply incompatible with this model, a model that is optimized mainly for biotech and B2B SaaS.

Investment in deep tech (think fusion energy, quantum computing, artificial general intelligence, autonomous robotics, etc.) requires tolerance to much longer timelines. This leads to a whole host of problems, not the least of which is forced liquidity of promising companies long before they have achieved their full potential. Patient capital is scarce in Canada. Deep tech companies have problems at both ends, struggling to find support to get off the ground, only for the small number that do to run into issues when their development timeline is mismatched from that of their investors.

The Deep Tech fund had timeline problems on two fronts. The first was structural: while 12 years is better than 10, the extra two years would not changed much in the long run. The 4 year extension would have helped, but even then, putting a hard time limit on truly disruptive technology development is always going to run into exceptions. How long have we been working on quantum computing, at this point? Second, the fact that the fund was shut down 4 years into its operational lifespan means we will never find out if that assertion is true, and reflects a fundamental issue with BDC as the parent organization for a fund that aims at creating long-term impact.

If BDC hosts another deep tech fund as their press release indicates, it will be critical to build in a tolerance for low liquidity on much longer timelines even than the first iteration intended, ideally removing the lifecycle requirement entirely, along with removal of any ability for defunding by the parent organization in response to external pressures like election cycles and unpredictable neighbors. Deep tech requires holding investments through the full range of market conditions and across political regime changes. Short term noise cannot play a role in decision making.

In any other investment context, an LP pulling out or failing to come up with capital when their commitment is called upon results in complete forfeiture of their entire stake in the fund. The consequences of failing to follow through on their commitment should be no less severe for BDC, if only for the sake of making such reactionary budget shifts not worth it.

Cheque Size

The BDC fund was designed to write cheques in the $5M range. In my view, this is too large to support emerging technologies by at least an order of magnitude. Canada’s premiere deep tech impact investors, Waterloo’s Velocity Fund, will tell you that most deep tech companies need about $1M to get off the ground and across the valley of death at pre-seed, and most of my research into deep tech funds tied to academic institutions supports this assertion.

Because of the oversized cheque, BDC’s Deep Tech fund had relatively few opportunities that fit the thesis, as most deep tech companies pre-revenue are raising rounds too small to be on their radar. Combined with the lack of deep tech investment in Canada generally, this means that companies have to survive the valley of death on their own before BDC was even able to get involved in supporting them further.

While the opportunities passed upon rarely form part of a fund’s performance metrics and is certainly not a dataset that will ever be made public, I suspect that many promising opportunities had to be turned down as a result of a mismatch between cash needed and the lower limit on BDC’s cheque size. This is symptomatic of the same risk-intolerance that plagues all of Canadian innovation. By insisting on a large cheque size, investments are only made after companies have made significant headway by other means, allowing for reduction of risk at the cost of many potential opportunities.

This is not to say that the $5M cheque size is not important, simply that it is not nearly as impactful as it could be without additional support earlier in the process. Deep tech companies need to raise Series A, too, but without a source of patient capital at a much smaller price point earlier in the pipeline, many companies will have failed, never launched, or left Canada long before the Deep Tech fund could have made a difference.

Any spiritual successor to the Deep Tech fund should aim for much smaller cheques, deployed much earlier in the process.

What Now?

Regardless of the performance of BDC as an organization or any structural issues in the first iteration, loss of the Deep Tech fund is a major blow to Canada’s deep tech ecosystem, as they were widely viewed as an anchor in the deep tech investment landscape. I expect that their absence will have a chilling effect on upstream investment by funds that previously considered them the most likely next investor in the chain, which Canada simply cannot afford.

If I am right in my assessment that the Deep Tech fund was just a casualty of broader efforts to course-correct at BDC, then it is clear that BDC’s risk tolerance is incompatible with deep tech generally, and BDC should ask itself some difficult questions about the unintended consequences of policy whiplash on an already fragile ecosystem before deciding that it is the right vehicle for delivering deep tech innovation.