
Estimated read time: 7 minutes
Canada has always been good at building the thing before the thing. The raw material. The foundational layer, or framework if you will. The quiet systems that someone else usually ends up scaling. We are a country full of early problem solvers – people who push the boundaries in research, who explore what’s possible, who seemingly get to the starting line first (in some cases). But somewhere between the idea, initiation, and impact, we fade away.
Apparent through a small amount of research – this happens across a multitude of sectors. In cleantech. In quantum. In robotics. In life sciences. In telecommunications. In dozens of technical and intellectual fields, Canada has played a quiet role. We invent, we test, we prototype. But when it’s time for the “impact” portion, the time to go big, to raise capital, to stake reputation, we look around the room and wait for someone else to go first. That is where the long-run upside disappears. People often like to say Canada has a commercialization, scale, and brain-drain like problem, however they miss one important variable… ambition. We as a society are structurally risk-averse. Our biggest players; government, big corporations, investors, and leaders – have grown to be too comfortable watching, rather than doing. Would rather be fast followers, than the early customer (big issue with our government).
Double click into AI
At this point I’d like to openly state that I’m a very curious and also very opinionated 22 year old with zero academic background in Computer Science.
Let’s continue. Nowhere is this pattern more clear than in artificial intelligence. On paper, Canada looks like a global leader. Our fingerprints are everywhere. Geoffrey Hinton helped define backpropagation and neural nets. Yoshua Bengio pushed the field of generative models forward. Richard Sutton literally wrote the textbook on reinforcement learning. The 2017 Transformer paper, which ignited the large language model wave, was co-authored by a Canadian undergrad at the University of Toronto. At this point it’s fair to say we aren’t simply tagging along on this new age of technology – we helped start it.
But. Within that opportunity of a lifetime to pave a new path for our economy, the chance to turn invention into infrastructure, research into revenue, the value moved. Quietly.
Element AI raised over $250 million with the ambition to turn Canadian AI into deployable enterprise tools. It had the top-tier talent, brand power, institutional support. But it was sold to ServiceNow (California), supposedly for what many considered to be a big discount. Maluuba, a deep learning company out of Waterloo with a “natural language focus”, was acquired by Microsoft (Washington) before most people even heard of it. CentML, working on model optimization, was acquired by Nvidia (California). Untether AI? Scooped up by AMD (California). In each case, the same story – built here, funded partially with Canadian dollars, sold off, scaled elsewhere.
Now, I know I’m taking a judgemental point of view on all of this. Truth is these guys got paid, and big time. They weren’t failures, they were technical success stories. The unfortunate truth is our ecosystem didn’t make it worth staying. We didn’t have the runway, the large scale capital, or the domestic buyers. So the moment these companies hit that inflection point, there was only one door open, and it wasn’t in Canada.
Deeper rooted issues in Canada
By now you’ll see the perspective I’m taking on this issue. How about I rifle off some more.
Our SR&ED tax credit system, while it looks generous on paper, is complex, slow, often doesn’t reward late-stage scaling the way it rewards early-stage experimentation. The trend in Canada seems to strand from a Federal and Provincial level – favoring pilots rather than persistence and longevity.
There is no incentive for our biggest institutions to be early customers of Canadian AI. No mandate. No urgency. So they don’t. The governments wait, the corporations wait – and startups, by definition, cannot afford to stand around and wait.
The legal environment seems to be adding friction as well. Cross-border regulations are meh. Privacy standards are high, which is a good thing but our enforcement and guidance feels seemingly ambiguous. This leads to founders spending more time and money to gain legal clarity in areas where the U.S. counterparts are already on the ground and running.
Then there’s capital. Our VC market is very cautious. We’re solid at pre-seed, and Series A, but are non existent beyond that. If you’re trying to raise a $100M round to build infrastructure, or train a model – odds are you’ll have to go south. Which means the board moves south, the IP goes south, and the talent ultimately goes south. A caveat to note: Yes, the U.S. has more capital, no question. But even if you normalize for scale, relative to one-another, their deal participation percentage dwarfs ours. As a share of Series B & C funding in Canada, domestic VCs are barely in the room. We’ve effectively trained a generation of founders to view scaling as something that happens elsewhere. And in doing so, we’ve exported not just talent, but control.
So yes, we have an ambition problem, a commercialization problem, a scale problem, and a brain drain problem.
Something is changing…
Here is the part that keeps me from closing the book on all of it: somethings changing. Not dramatically, but in conversations and through people deciding they’re tired of all the good in Canada leaving and calling it normal.
One of these signals is a company called AXL– a new venture studio in Toronto that isn’t trying to be flashy, but necessary. I’ve seen first hand, having met them, that they aren’t simply talking about taking it to the next level but actually designing around the gaps in the current system. They are building with researchers, not after them. Starting with deep customer problems, not just building prototypes in search of a market. It’s being led by people who’ve been inside the machine, people who have built, exited, raised capital, worked with big names like Meta, Telus, etc. At AXL, they’re choosing to build 50 Human Centric AI companies here in Toronto, and that sends a message to people like me – that there is a path here and maybe the default shouldn’t be going south, but instead building here.
Yet still, I hesitate. Because one amazing group of people may not be enough. A promising project may not solely patch decades of passivity. If we want to actually keep what we create, we need the government to get serious about being a first customer – not a follower. We need our banks, telco’s, and funds to stop acting like risk is someone else’s jurisdiction.
Additionally, an uncomfortable truth: most of the young people I know don’t actually want to leave. They just don’t see a future or enough proof that staying here in Canada leads anywhere. I’m 22. I think about this stuff more than I probably should. But when you grow up watching your country get credit for all of the ideas and retain none of the value, you start looking at things through a different lens. It feels too similar to a participation badge in one big game of business and life.
The talent is here. The research is here. The ideas are here.
The only question left is whether we have the balls to hold onto any of it.
Prediction
Over the next decade, Canada will be forced to confront a deeper question than “how can we keep our talent?”. The real shift will come when we’re no longer judged by the research and breakthroughs we produce, but by the IP and ideas we keep. In a world where talent is more mobile, and ideas are less bound to geography, the countries that succeed will be the ones that create immense value and meaning from the funnel of IP they can generate and retain. If Canada doesn’t build out a longevity focus and the long end of this funnel – culturally, economically, structurally – we won’t just continue to lose our best people to Silicon Valley and big tech rollups. We’ll miss the far bigger opportunity to be something better than we are now.
Nick Trent
Leave a comment