Canada is moving into the AI race with a strategy that goes beyond grants, training programs and regulatory language. The federal government is setting up a C$500 million Canadian Tech Growth Fund, roughly US$360 million, to support promising domestic artificial intelligence companies. The fund is designed to provide flexible capital, help companies scale, attract private investment, retain intellectual property and keep more of Canada’s AI industry anchored at home.
The most important detail is not only the size of the fund. It is the structure. Ottawa may take equity stakes in some of the country’s most promising AI companies.
That changes the conversation. AI policy is usually framed around risk, safety, adoption, competition or national security. Canada’s approach adds an ownership question: if public capital helps create the next generation of AI companies, should the public receive a direct claim on the upside?
The answer matters because AI is not developing like an ordinary software market. It requires enormous amounts of capital, compute, data, talent and infrastructure. Countries that cannot finance their own AI firms risk becoming dependent on foreign platforms for the systems that will shape work, healthcare, education, public services and business productivity. Canada’s strategy is partly a response to that dependency. The government has warned that reliance on foreign suppliers for compute, cloud infrastructure and data storage creates strategic risks, including exposure of Canadian data and weakened domestic competitiveness.
The new fund sits inside a broader national strategy called “AI for All.” The government wants to increase AI adoption among Canadian businesses, expand sovereign compute capacity, invest in public-interest AI missions such as healthcare, strengthen AI safety research and provide AI literacy training. Those goals make the strategy broader than industrial policy alone. It is also a bet that AI can raise productivity and create new firms, jobs and public value.
But the ownership question is what makes the fund especially relevant. Public investment in technology often follows a familiar pattern: taxpayers absorb early risk, private investors capture most of the upside, and the public later pays again to regulate, procure or repair the consequences of the technology. An equity stake offers one way to change that pattern. If the state helps finance AI companies at a critical stage, it can potentially share in future gains.
That does not automatically make the model democratic or fair. Public equity can become passive financial participation without broader accountability. The government may own shares, but workers, users and communities may still have little voice in how AI systems are deployed. A public stake can also raise difficult questions about political influence, investment selection and whether the state is picking winners.
Still, the move is important because it challenges the idea that public support should be limited to subsidies. If AI is going to produce large private fortunes, reshape labor markets and rely on publicly supported infrastructure, then the public has a legitimate claim to more than indirect benefits. Ownership gives that claim a financial form.
The same question applies to workers. Canada’s strategy aims to create jobs and raise adoption, but AI will also reorganize work. If productivity gains flow mainly to shareholders, vendors and executives, the technology will deepen the divide between those who own productive systems and those whose labor is transformed by them. Public equity stakes are one answer at the national level. Employee ownership, profit-sharing, co-operative data governance and public-interest AI funds are related answers at the firm and community level.
The strongest version of Canada’s approach would connect these layers. Public capital could help keep AI firms domestic. Equity stakes could return value to citizens. Compute investments could reduce dependency on foreign platforms. Safety and privacy rules could protect users. Worker and community ownership models could ensure that productivity gains are not captured only at the top.
That is the deeper test. Sovereign AI cannot only mean Canadian companies rather than foreign companies. It has to mean a different distribution of power around the technology. Who owns the firms? Who controls the infrastructure? Who governs the data? Who benefits when productivity rises?
Canada’s AI growth fund is a signal that governments are beginning to understand AI as an asset question, not just a regulation question. The next step is deciding whether public investment will simply help build national champions, or whether it will also build a broader ownership stake in the economy those champions create.