According to Stanford University, there were 1,953 newly funded American AI companies in 2025 versus 79 Canadian ones. Photo by ANDREJ IVANOV/AFP via Getty Images files

Article content

Around the world, economists, technologists, politicians and philosophers are debating AI’s likely transformative yet highly uncertain potential for good and ill. As Canada grapples with challenges from a debilitating trade war and self-inflicted economic weaknesses, decision-makers have a historic opportunity to harness AI in the national interest.

Article content

Article content

One critical issue is whether AI can rescue Canada from 10 years of productivity stagnation resulting from falling Canadian real investment per worker, which was down a grim 15 per cent between 2014 and 2023 — compared to up 21 per cent in the U.S., 13 per cent in the Euro area and 11 per cent in the OECD.

Article content

Article content

Article content

Also critical is whether Canada’s regulatory, energy and investment policies will allow the country to exploit its early AI research advantage, given our historic weakness in converting innovation into higher productivity and global-scale companies.

Article content

Canada was one of the birthplaces of modern AI, with crucial work done by Geoffrey Hinton in Toronto, Yoshua Bengio in Montreal and Richard Sutton in Alberta. Yet the AI giants are overwhelmingly American: Nvidia, Apple, Google, Microsoft, Meta, OpenAI and Anthropic. Canada’s private AI investment gap with the U.S. last year was enormous — US$286 billion versus US$4.3 billion, or 67 times more in an economy that is about 13 times the size of Canada’s.

Article content

Article content

Article content

Statistics Canada cites estimates about how much AI could increase annual Canadian labour-productivity growth over the next decade that range from 0.4 to 1.1 percentage points. It reports that in the second quarter of this year 19.2 per cent of Canadian businesses had used AI to produce goods or services during the previous 12 months, which is roughly comparable to U.S. numbers.

Article content

Article content

But adoption, capital investment and commercialization have been limited. According to the Bank of Canada, more than two-thirds of business leaders personally use AI during a typical week, yet only eight per cent apply it significantly in their core operations, while another 50 per cent report only low or moderate core use. Really large economic gains could come from enhancing productivity in energy, mining, construction, manufacturing, transportation, agriculture and health care, where AI adoption is currently much lower.

Article content

The government identifies Canada’s cold climate and hydro, nuclear and other electricity resources as competitive advantages in AI. But we need vastly more generation, transmission and other infrastructure to meet projected demand for 5.5 GW of AI compute by 2030, which is only a little over three years away. Ottawa says it wants to double the electricity grid by 2050, but it is not at all clear our regulatory and permitting systems can deliver that much new infrastructure in a timely manner.