The recent release of DeepSeek’s R1 artificial intelligence model presents a stark illustration of what Canada has lost – and what it could regain. This Chinese lab has achieved performance comparable to OpenAI's leading models at roughly 1/27th the cost, demonstrating that breakthrough AI development no longer requires bleeding-edge hardware or billion-dollar budgets.
This development should serve as both a wake-up call and an opportunity for Canadian policy-makers.
The prevailing narrative in AI development has centred on massive computational requirements and corresponding capital expenditure. DeepSeek, however, accomplished its breakthrough without legal access to the latest NVIDIA H100 chips or the vast data centres of Silicon Valley. Instead, they succeeded through clever engineering and algorithmic innovation – exactly the kind of intellectual capital in which Canada once excelled.
Recent research has sadly shown a dramatic decline in Canada’s elite AI talent – from hosting 10 per cent of top researchers in 2019 to just 3 per cent in 2022. In light of DeepSeek’s achievement this failure to retain talent takes on even greater significance. The fact that only 37.5 per cent of top researchers who completed graduate studies in Canada post-Covid remained here, compared to half in 2019, suggests a deeper structural issue that mere capital investment won’t solve.
The solution lies not in matching the massive expenditures of American tech giants, but in fostering an environment that enables our researchers to stay here and innovate effectively.
The imperative for Canada extends beyond technological considerations – it speaks to our fundamental capacity for self-determination in an era of mounting existential pressures. DeepSeek’s achievement and enormous market impact demonstrates that when intellectual advances and capital efficiency combine, they become more than technological achievements – they become assertions of technological sovereignty and national capability.
Canada’s past successes in AI development were no accident; they resulted from creating a supportive research environment for our leading scientists and researchers. When Geoffrey Hinton, the British-Canadian “Godfather of AI,” sought funding for neural networks research in the 1980s, he encountered an American research ecosystem thoroughly captured by military imperatives. The Canadian Institute for Advanced Research, by contrast, embodied a distinctly Canadian approach to knowledge creation – one that recognized the inherent value of pure research.
In an era where talent retention has become inextricably linked with national sovereignty, Canada must articulate a compelling vision that connects our historical achievements to our future aspirations.
Canada’s recent $2-billion capital investment in AI compute infrastructure, while important, needs to be complemented by a broader strategy that recognizes what truly drives innovation – people.
Rather than focusing on investing primarily on hardware and computation – where we can’t hope to compete relative to the US – we should emphasize policies that make Canada an attractive destination for AI talent. DeepSeek has shown that this is possible. Canada already has a competitive advantage in our established pedigree in AI innovation that needs to be rebooted and revitalized. To attract talent, we need competitive tax rates, a reversal of our proposed capital gains tax hike on innovators, establishing regulatory sandboxes for AI applications in critical fields such as energy and health care rather than the onerous proposed Artificial Intelligence and Data Act requirements, and increased support for commercialization of research. Simple changes, like fast-tracking immigration for AI researchers and their families, or increasing/streamlining SRED (Scientific Research and Experimental Development) for AI startups, could have outsized impacts. We must focus on creating an environment that naturally attracts and retains top talent.
As the need for AI talent heats up, the US is clearly aiming to lure more of our top talent, and with a more attractive financial landscape and a greater variety of tech companies to work for, it already has a leg up.
This moment calls for more than policy adjustments – it demands a philosophical rethinking of how we develop and deploy technology. The threat of tariffs and pressure on our sovereignty can serve as a catalyst for reimagining how we leverage our technological capabilities.
Canada can build upon its past successes in AI. But we must act decisively to create an environment where innovation can flourish. The alternative is watching other nations continue to leverage our intellectual legacy while we fall further behind.
Ryan Khurana is a senior fellow at the Foundation for American Innovation and a contributing author to the Macdonald-Laurier Institute.