The Government of Canada is investing $42.5 million to deploy new artificial intelligence (AI) computing infrastructure at the University of Toronto (U of T) that will be used by researchers across the country.
This funding will expand U of T’s existing AI compute capacity to support AI research from coast to coast. Canada’s Minister of AI and Digital Innovation Evan Solomon shared the news today from Toronto AI research hub the Vector Institute. Solomon said the funding will support training big AI models and provide a secure, multi-tenant platform for data-intensive AI research.
“Countries that control their own compute will have control of their own future.”
“This is an investment in our sovereignty,” Solomon said on stage. “This is an investment in keeping Canadian data here in Canada. This is an investment in Canadian discoveries powered by Canadian infrastructure … Part of this is what our AI strategy is all about. Countries that control their own compute will have control of their own future.”
Canada’s underinvestment in computing power has threatened its AI talent and research advantage. The federal government’s $2-billion Sovereign AI Compute Strategy aims to help rectify this. This financing for U of T marks the second major commitment through that strategy after the feds last year promised Toronto large-language model maker Cohere up to $240 million to purchase capacity at a new data centre in Cambridge, Ont.
“The practical takeaway is, we’re going to be able to train and validate these [AI] models that have billions of parameters, and this was not possible before in the Canadian academic ecosystem, and we want to make this accessible to anyone,” Timothy Chan, associate vice-president and vice-provost of strategic initiatives at U of T, told BetaKit in an interview.
The funding has been allocated via the Digital Research Alliance of Canada, a federally-backed non-profit that provides digital tools and infrastructure to domestic researchers. This $42.5 million for U of T includes a $40-million capital investment in the 2025–2026 fiscal year and $2.5 million over the following two years to operate, maintain, and staff the facility.
RELATED: AI minister won’t rule out American involvement in Canada’s sovereign cloud
U of T plans to invest an additional $100,000 in the initiative. The university’s supercomputing centre SciNet is overseeing the system, which U of T says will complement its existing capabilities and let researchers leverage AI for a variety of applications, from healthcare to fundamental science and tech.
Chan told BetaKit that the $42.5-million investment will largely go towards the purchase of new graphics processing units (GPUs), the chips that power AI, from American giant Nvidia, with a smaller fraction dedicated towards increased storage.
These GPUs will be co-located with U of T’s existing Trillium system, a high-performance computing cluster that the university owns and SciNet operates, which is housed at a data centre in Vaughan, Ont. Launched this August, Trillium is one of the country’s most powerful supercomputers for academic research. Chan expects this federal investment to triple that system’s GPU computing capacity. He noted that U of T and SciNet are already working on this expansion, which is expected to be complete by spring 2026.
The feds are also working to set up the up to $705-million AI Sovereign Compute Infrastructure Program (SCIP), which will support the establishment of a new, large-scale, Canadian-owned supercomputer to support researchers and industry, with a request for proposals to come in 2026.
RELATED: Nokia’s Ottawa expansion tests Canada’s definition of “sovereign AI”
While this investment at the U of T contributes to Canada’s sovereign computing goals, Solomon has argued that a wide variety of partnerships—including those with foreign multinationals—are needed to achieve that strategy. Earlier this week, Solomon described Finnish telecommunications giant Nokia’s Ottawa expansion—a $340-million project that the feds are providing $40 million towards—as “what sovereign AI looks like.”
When asked how he responds to Canadian tech leaders like Shopify co-founder and CEO Tobi Lütke calling the government’s involvement “toxic” and Council of Canadian Innovators president Ben Bergen arguing that cutting cheques to foreign multinationals does not “build sovereignty for Canada,” Solomon said the feds want companies to invest in Canada, and argued that there is “room for both” successful Canadian-owned businesses like Shopify and international players looking to grow here like Nokia.
“Nokia has been here for a long time,” Solomon told BetaKit during a media scrum following his prepared remarks. “They just built a $340-million data innovation hub in Canada that’s going to provide another 340 jobs … It is excellent to see companies who are investing hundreds of millions of dollars in Canada.”
Amid a United States trade war where the feds have prioritized ensuring Canada’s resilience and building deeper relationships with other international trade partners, Solomon argued, “it is always surprising when a company chooses Canada like this, and this is a good story.”
Feature image courtesy the University of Toronto. Photo by Lisa Lightbourn.