The AI Buildout Needs Talent and Clean Power. Canada Has Both.
Canada pairs a decade-old AI research base with a $2-billion sovereign compute strategy, hydro-heavy power and a tech workforce that costs less than half of Silicon Valley's.
Every serious AI investor now runs the same two-line arithmetic: where are the people who can build the models, and where is the power to run them. Most jurisdictions can answer one line. Canada answers both, and it did so before the current buildout began.
The research base is not a 2026 improvisation. The federal government funded the Pan-Canadian Artificial Intelligence Strategy with $125 million in 2017, then added $443.8 million in Budget 2021 and a further $174 million in the 2024 Fall Economic Statement. That money anchors three national institutes, Mila in Montreal, the Vector Institute in Toronto and Amii in Edmonton, together with a ten-year, $208-million commitment to CIFAR for talent and research. On May 21, 2026, Ottawa and CIFAR announced $24 million to appoint or renew 42 Canada CIFAR AI Chairs, bringing the national total to 143. The same institutes graduate more than 250 graduate students a year, which is the pipeline the rest of this argument depends on.
The commercial layer has caught up with the academic one. Toronto-based Cohere closed a US$500-million round in 2025, then a US$100-million extension that pushed its valuation to US$7 billion, with the Business Development Bank of Canada, PSP Investments and HOOPP on the cap table beside NVIDIA, AMD Ventures and Salesforce Ventures. Its annualized revenue was US$100 million in May 2025 and was projected to pass US$200 million by year-end.
Compute: a $2-billion strategy that is now disbursing
Talent without compute is a paper strength, which is why the Canadian Sovereign AI Compute Strategy matters. It commits $2 billion over five years from 2024-25 across three lines: up to $700 million to mobilize private data-centre projects, up to $1 billion for public supercomputing infrastructure, and up to $300 million for an AI Compute Access Fund that pays two-thirds of eligible compute costs when a small or mid-sized company buys Canadian cloud capacity, and half when it buys non-Canadian capacity.
The fund is no longer a promise. On May 12, 2026, at Web Summit Vancouver, Evan Solomon, the Minister of Artificial Intelligence and Digital Innovation, named 44 companies sharing $66 million in the first round, in sectors from wildfire detection to transit routing, and described the program as oversubscribed. The day before, Ottawa and TELUS signed a memorandum of understanding on a large-scale sovereign AI data centre in British Columbia under the government’s call for proposals that ran from January 15 to February 15, 2026. TELUS chief executive Darren Entwistle said demand had “completely sold out” the company’s first AI facility in Rimouski, Quebec. No federal money has yet been committed to the BC project, and the government says so plainly; the signal is that the pipeline exists and the private counterparty is a listed utility, not a start-up.
The provinces are moving on their own clock. Alberta has set a target of $100 billion in data-centre investment by 2030, and its technology minister, Nate Glubish, said in June that “gigawatt-scale announcements” would follow later this summer, with the first campus 12 to 18 months from switch-on. His condition to developers was blunt: “If you cause the cost on our grid, you’re going to pay for it.” That is the kind of user-pays discipline a long-term investor should want to hear.
Power: the hydro dividend
The reason those campuses can be sited in Canada at all is the electricity mix. The Canadian Centre for Energy Information reports that 57 per cent of Canadian electricity came from hydro in 2023, with nuclear adding about 14 per cent and 80 per cent of generation coming from non-emitting sources. Statistics Canada found that 83.3 per cent of the $11.8 billion spent on new power plants in 2024 went to non-emitting sources, and it projects wind and solar capacity to double by 2035 as national energy needs grow 62 per cent by 2050. A hyperscaler that has promised its shareholders low-carbon compute cannot easily buy that in Texas or Virginia. It can buy it in Quebec, British Columbia, Manitoba and, increasingly, Ontario.
The labour arithmetic
CBRE’s Scoring Tech Talent 2026 report, published August 18, ranks Toronto third in North America with 33,419 AI jobs, Vancouver ninth and Montreal eleventh, up four places in a year. The cost gap is the investor’s number. Average annual tech wages run US$211,048 in the San Francisco Bay Area and US$190,050 in Seattle against $70,185 in Edmonton and $71,977 in Quebec City. A 500-person tech operation costs about $36.1 million a year in Quebec City versus $90.6 million in the Bay Area.
Immigration policy is being tuned to feed that pool. Express Entry already runs a STEM category, and Immigration, Refugees and Citizenship Canada’s 2026 consultation, which closed September 1 with more than 17,000 submissions, put two new priorities on the table for 2027: an accelerated pathway for H-1B visa holders, a commitment made in Budget 2025, and a Canada Global Impact+ Research Talent Initiative to recruit leading international researchers. For a company whose engineers are stuck in a US visa lottery, that is a concrete operating option, not a slogan.
The honest caveat is scale. Canada’s compute strategy is $2 billion; the four largest US hyperscalers, Amazon, Alphabet, Microsoft and Meta, have guided to roughly US$725 billion of capital spending in 2026 alone, most of it on data centres. Canada will not out-build them. It does not need to. Its pitch is narrower and more durable: the research institutions have been funded for nine years, the graduates come out every spring, the electricity is clean and priced, and the government has started writing cheques rather than strategies.
Why now: the first Compute Access Fund round has closed and ISED says applicants should watch for any new call, the sovereign data-centre MOUs are being negotiated this year, and the 2027 immigration categories are being decided this fall. The investors who arrive in 2026 negotiate the terms; the ones who arrive in 2028 accept them.
Sources
- Canadian Sovereign AI Compute Strategy (ISED)
- AI Compute Access Fund (ISED)
- Feds announce $66 million for 44 businesses through AI Compute Access Fund (BetaKit)
- Government of Canada and TELUS advance work to build sovereign AI infrastructure (Canada.ca)
- Pan-Canadian Artificial Intelligence Strategy (ISED)
- Government of Canada and CIFAR announce $24M investment in top AI talent (CIFAR)
- Cohere's valuation hits $7 billion USD following $100-million round extension (BetaKit)
- Rapid Growth of AI-Related Jobs Strengthens Top North American Markets, Scoring Tech Talent 2026 (CBRE Canada)
- 2026 consultations on economic priorities for category-based selection in Express Entry (IRCC)
- Alberta May See 'Gigawatt-Scale' AI Data Centres Under Construction This Year (EnergyNow)
- Clean power and low-carbon fuels, Energy Facts (Canadian Centre for Energy Information)
- Meta, Microsoft, Amazon, and Alphabet are about to spend a shocking amount of money to dominate the AI era (Yahoo Finance)
- Investment in non-emitting electricity generation rises as demand grows in Canada (Statistics Canada)