AI sovereignty isn’t the answer for Canada’s economic future

In recent months, Canada has been pulled into a global debate about how best to regulate and develop artificial intelligence. On one side are advocates for a global, innovation-first approach, emphasizing that Canadian innovators should seize the productivity opportunities AI promises for Canada. In this view, AI is the next “general-purpose technology”—akin to electricity or the internet—and should be unleashed to drive growth, wealth, and security. On the other side are voices calling for what they term “digital sovereignty.” Some are urging  Ottawa to prioritize “technological non-alignment” and heavy domestic control of AI infrastructure. Their arguments rest on the belief that if Canada does not “own” and “host” its AI systems domestically, we will become economically dependent on foreign-owned platforms, a modern version of the branch plant economy. This divide mirrors an older debate in Canadian economic history: whether the country thrives by embracing openness or by retreating inward. And history offers a clear verdict: when Canada chooses openness, Canadians prosper. This lesson matters now more than ever as the country confronts stagnating productivity and slow GDP growth. If AI is indeed the transformative technology of our time, the wrong approach could lock us into a lost generation of underperformance. What’s more, the promise of AI, and open source in particular, offers a chance to move beyond economic debates of the past, allowing governments to harness collaborative technologies securely, in the national interest. Canada’s failed history with economic nationalism Economic nationalism has deep roots in Canada. From the early 20th century through the 1970s, Canadian policymakers frequently sought to shelter domestic firms from foreign competition. They invoked slogans about sovereignty, independence, and protecting Canadian identity. Yet again and again, these experiments produced disappointing results: lower growth, weaker innovation, and higher costs for ordinary Canadians. Waves of economic nationalism promised to secure Canadian independence and prosperity, but each left the country poorer, less innovative, and more dependent on foreign technologies developed elsewhere. By contrast, when Canada has embraced openness—through the Auto Pact, the Canada-U.S. Free Trade Agreement, and NAFTA, for example—the economy has flourished. In each case, access to global markets spurred productivity, investment, and job creation. Canadians benefited from lower prices, greater choice, and rising living standards. Today, Canada faces what economists and commentators describe as a “growth crisis.” Per-capita GDP has stagnated for a decade, and the OECD projects Canada will post the slowest per-capita growth among its members by 2060. This is not a time for nostalgic experiments in protectionism. It is a time to double down on openness, integration, and policies that foster investment and innovation. Why pursuing AI sovereignty would be a costly mistake  Despite this history, some argue that AI is different, that Canada must host and own its AI systems domestically to protect national interests. They warn of “foreign infrastructural capture,” and propose policies ranging from mandatory data localization to public investment in Canadian-only platforms. But while efforts like “Buy Canadian” make sense to reduce trade barriers between provinces, for example, the broader rhetoric that accompanies them masks a familiar agenda: inefficient public investment in areas where governments have little expertise.  It’s digital-era branding for plain protectionism. There are several reasons why pursuing AI sovereignty would be misguided. First, AI is not like cultural content or even energy. It is a general-purpose technology closer to an intermediate good in global trade. AI models, cloud services, and large-scale compute are inputs into countless other industries—from health care to logistics to manufacturing. Restricting Canadians’ access to the best and cheapest AI tools is equivalent to imposing tariffs on machinery for advanced manufacturing. It raises costs, reduces competitiveness, and slows innovation. Second, Canadians already benefit from global AI investment. Pension funds, mutual funds, and institutional investors hold significant stakes in U.S. and global technology firms, driving AI progress. The Maple 8, for example, has invested nearly $6.6 billion USD into publicly traded, U.S. digital infrastructure and data centre firms. Ordinary Canadians’ retirement security is tied to the success of companies like Microsoft, Nvidia, and Alphabet. To shut out these firms is to misunderstand how integrated our prosperity is with global markets in the past, present, and future. A Canadian-branded alternative would not change the underlying economics of AI development; it would simply diminish the benefits to Canadians and narrow the options for Canadian businesses, while duplicating costs to deliver less performant tools. Third, protectionism would carry real economic costs. Recent studies found that data localization rules alone could shave up to 1 percent off GDP and increase data hosting costs by 30 to 60 percent. Open source models democratize generative AI technology, making it accessible to all, regardless of access to computing power or technical expertise, or ability to secure funding to create foundation AI models. Llama models from Meta, for example, are freely available to individuals, governments, and businesses in Canada to harness for national advantage. These are not abstract figures or concepts. Blocking access to global technology translates into higher costs for Canadian businesses and consumers, fewer opportunities for small firms to adopt AI, and slower diffusion of productivity gains across the economy. Fourth, national security concerns are overstated. Advocates of digital sovereignty portray U.S. platforms as tools of geopolitical control, but these are profit-driven firms focused on market share, ads, and cloud contracts. Canadian businesses using Google Cloud or Microsoft Azure are making rational cost and performance choices, not surrendering sovereignty. Data security can be safeguarded through open-source models, which run locally and allow Canadians to build applications tailored to local needs while keeping data in Canada. Ultimately, Canada’s security rests on allied integration, not self-sufficiency in AI infrastructure. Finally, Ottawa lacks the state capacity to deliver on AI sovereignty. The federal government has enough on its plate with basic service delivery. Meta has invested billions of dollars to deliver high-performing AI models and improves those models continuously. Microsoft and Amazon, similarly, have been building global cloud capacity for decades. To imagine that Ottawa could—or should—design and operate globally competitive AI infrastructure, regulate platforms, and direct investment at scale, maintaining the race to build new foundational models, is simply not realistic. The likely result would be vast public expenditures, underperforming domestic platforms, and frustrated users who turn to global alternatives anyway. The irony is that Canada already has a genuine comparative advantage in AI research. Our universities and research institutes, particularly in Toronto, Montreal, and Edmonton, are globally recognized hubs. But instead of leveraging this head start into open integration with global AI ecosystems, Ottawa risks squandering it on nationalist detours. The future of AI in Canada should be about maximizing diffusion, helping firms across sectors adopt AI tools to boost productivity, not about isolating infrastructure in the name of sovereignty. AI, like electricity or the internet, will shape the next century of growth. Canada cannot afford to repeat the mistakes of the NEP or FIRA. Closing ourselves off would guarantee that we fall further behind our peers. Embracing openness, by contrast, offers a path to renewed dynamism. Key takeaways: Learning the right lesson The debate over AI adoption in Canada is not merely technical. It is a test of whether we have learned from our economic history. Every time Canada has chosen economic nationalism, it has lost. Every time it has embraced openness, it has thrived. The debate is also a false choice. By adopting AI tools on freely available open source platforms, governments can benefit from globally competitive and integrated technologies, working securely in the national interest. Laissez-faire AI optimism and Canadian calls for “digital sovereignty” may seem like polar opposites, but they reflect the same question: do we view AI as a threat to be managed or as an opportunity to be seized? Canada cannot afford to get this wrong. AI sovereignty would saddle Canada with higher costs, weaker innovation, and slower growth—exactly the outcomes we can least afford amid today’s productivity crisis. The right path is to double down on openness: foster competitive regulatory conditions, integrate into global AI supply chains, and help Canadian businesses adopt the best tools available, wherever they are developed. If Canada takes this path, AI can become the technology that helps lift us out of stagnation and secures prosperity for a new generation. If we choose nationalism instead, we risk condemning ourselves to another lost decade. The choice should be clear. This article is presented in partnership with Meta.