A smart-sounding dumb idea
The Carney government’s plan for a public AI supercomputer is the kind of idea that sounds bold and strategic until one pauses to think about it for more than a few seconds.
The premise, championed by AI Minister Evan Solomon, is that Canada needs sovereign computing power to compete in the AI race. But the proposal is less a serious economic policy than a symptom of a broader and worrying shift toward renewed economic nationalism.
Consider the scale of the industry in question. The global AI leaders—Anthropic, Google, Meta, OpenAI and others—are spending hundreds of billions of dollars annually on chips, data centres, energy infrastructure, and cloud capacity. Their yearly capital expenditures rivals or exceeds Ottawa’s total budget.
Against that backdrop, the idea that the federal government will build and operate a globally competitive AI supercomputer borders on absurdity. This is, after all, the same government that spent years failing to implement a functioning payroll system. Canadians are now supposed to believe it can successfully manage one of the most capital-intensive and technologically dynamic sectors in the world.
Yet this proposal matters beyond its own merits. Before Donald Trump’s election in November 2024, Canada’s secular stagnation seemed to be producing a healthier policy conversation. There was growing recognition that weak productivity and low investment required more competition, greater openness, and structural reform. Even previously untouchable issues, such as foreign ownership restrictions in protected sectors, were beginning to face scrutiny.
Trump’s repeated attacks on Canada have changed the mood. They have created political space for a new sovereignty economics: subsidized supercomputers, state-backed satellite launch capacity, and other industrial policies justified in the name of national independence rather than market dynamics.
That would be a costly mistake. Canada’s main economic problem isn’t that the state owns too little. It’s that the private economy invests too little, builds too slowly, and faces too many barriers to growth.
We don’t need a public AI supercomputer. We need a more competitive economy.