· Solveion · Perspectives · 5 min read
We never needed to
Canada did not miss the AI revolution. Canada started it, and the value settled somewhere else. As a Canadian firm we have been asking why, and the most honest answer is uncomfortable: for a long time there was no reason to keep it.
We are a Canadian consultancy, and like most people here we have spent this year thinking about a question that used to feel abstract. Canada has the ingredients of an economic superpower — resources, education, stable institutions, an enviable reputation, and an unusual ability to attract people from everywhere. So why are we not one?
The numbers have got harder to look away from. Canadian GDP per capita fell from 83.1 per cent of the American figure in 2014 to 71.4 per cent in 2024. In absolute terms that is roughly $54,400 against $85,300. Statistics Canada put the widening of the productivity gap at 26 per cent since 2000. Not since 1945 has the United States been this far ahead.
There are the usual explanations — a small domestic market, a resource-heavy economy, capital that prefers real estate to risk. All true, all incomplete. We think there is something underneath them, and it is not a failure of talent or ambition.
The comfortable assumption
For most of living memory, Canadians have not really thought of the Canadian economy as a separate thing that needed to win. We thought of a continent that we were part of, run by people we considered friends, in which prosperity was broadly shared and the question of whose column it landed in was a technicality.
That was not naive and it was not a lack of patriotism. It was a reasonable read of a genuine partnership. But it had a quiet consequence: if a Canadian invention scaled in California, that was not a loss. It was the family business doing well. There was no urgency to build the domestic capacity to keep it, because keeping it was not obviously the point.
We never needed to. That is the whole explanation, and everything else is downstream of it.
AI is the clearest evidence
Here is the part that should be uncomfortable for anyone who thinks this is a story about Canada lacking capability.
Canada did not miss the deep learning revolution. Canada is substantially where it came from. Geoffrey Hinton’s work at Toronto, Yoshua Bengio’s at Montreal, Richard Sutton’s at Alberta — three of the field’s foundational figures, holding Turing Awards and a Nobel between them, all doing that work here, much of it funded when nobody else thought it was worth funding.
The intellectual claim is about as strong as a country can have. The economic claim is not. The market value created on top of that research sits overwhelmingly in the United States, and the institutes that produced it — Mila, Vector, Amii — now depend substantially on foreign compute to do their own work.
The Walrus put the question more sharply than we would have dared: Canada is spending billions on AI, so why are companies still leaving? That is the right question, and it is not a question about talent.
What is actually missing
Not ideas. Not people. Not reputation, which may be Canada’s single most underrated asset in a period when trust between countries has become scarce and expensive.
What is missing is the machinery that turns invention into ownership. Canadian business spent about $9.3 billion on R&D in 2024, which puts us 21st of 44 countries as a share of GDP, at roughly 0.4 per cent. Investment here has flowed toward real estate and resource extraction rather than the sectors that compound. And on the specific input that matters most for this frontier — compute — we have been renting from other people’s countries.
None of that is destiny. All of it is the accumulated result of decades in which building our own was optional.
Why AI, and why now
We would argue AI is the right place to start, for three reasons that have nothing to do with fashion.
It is where economic weight is about to be redistributed, which makes it one of the rare moments when position is up for grabs rather than inherited. It is the one frontier where Canada has a genuine founder’s claim rather than a catch-up story. And self-reliance in AI is not self-contained: it propagates into healthcare, energy, agriculture, manufacturing, financial services and government, so capability built here compounds in a way that capability rented elsewhere does not.
Canada also has advantages that are easy to overlook because they are not technical. We can raise capital internationally from people who are wary of the alternatives. We can recruit globally from a position of genuine attractiveness. We have clean energy and cold weather, which is not a small thing when you are siting data centres. And we have the standing to build alliances that are not available to larger, louder countries.
The strategy work is under way — sovereign compute, national investment, a build-partner-buy posture. Our honest reading is that it is directionally right and still too slow, too fragmented, and too modest for the size of the opening.
The point is not to stop being friends
We want to be careful here, because it would be easy to read this as a turn inward, and that is not what we mean.
Closeness with the United States has been genuinely good for this country, and we would like it to continue. The argument is narrower: “our prosperity and theirs are the same thing” was a decision, not a law of nature. It was a good decision for a long time. It made building our own capability feel unnecessary, and now the assumptions underneath it have visibly moved.
A country with this much talent, this much credibility and this much runway should not be renting the foundations of the next economy. Not because anyone wronged us, but because we finally have a reason to do the thing we were always capable of doing.
We never needed to. That is no longer true, and we think it is the most useful sentence any Canadian organization could sit with this year.