U of T AI Infrastructure Gets $42.5M Federal Boost
Canada is investing $42.5 million to expand artificial intelligence compute at the University of Toronto, giving researchers across the country more capacity to push health, science, engineering, and the humanities forward.
The funding comes via the Canadian Sovereign AI Compute Strategy and will be delivered through the Digital Research Alliance of Canada's National AI Compute - Rapid Deployment initiative. The goal is direct: meet surging demand for high-performance AI computing and keep Canadian research on Canadian infrastructure.
Why it matters
Researchers have been leaning on foreign compute to train large models and run data-heavy experiments. The new system, operated by U of T in collaboration with SciNet, is built to support training at the scale of hundreds of billions of parameters and provide a secure, multitenant platform for data-intensive workflows.
This builds on last year's $52-million upgrade to the SciNet supercomputer at U of T, expanding national capacity for AI workloads and advanced simulations.
Access will extend beyond U of T. Small universities, research hospitals, northern and Indigenous communities, and industry partners will be able to tap into the infrastructure through national channels.
What researchers can expect
- National AI compute, operated locally: U of T and SciNet will run the system, focusing on secure, shareable resources for large-scale model training and data-centric research.
- Funding schedule and support: $40 million in 2025-26, plus $2.5 million over the following two fiscal years for talent and operations. U of T will add $100,000, while SciNet provides technical expertise and ensures environmental and operational best practices.
- Broader impact areas: Drug discovery, climate research, engineering, and computational work in the social sciences and humanities will all benefit from larger, faster workloads.
- Stronger data stewardship: Keeping compute and research data in-country supports sovereignty, security, and collaboration at scale.
Leadership and context
U of T has long been a leader in AI research. University Professor Emeritus Geoffrey Hinton's deep learning breakthroughs laid the groundwork for modern AI and earned him the 2024 Nobel Prize in Physics. Alumni and faculty continue to push the field forward, including Ilya Sutskever (OpenAI co-founder; co-founder and chief scientist at Safe Superintelligence), Sanja Fidler (NVIDIA), and Raquel Urtasun (founder and CEO of Waabi).
The investment underlines a coordinated approach between academia, industry, and government to accelerate discovery, develop talent, and keep Canada competitive.
The bigger picture
Public compute and research data are national assets. With strong AI infrastructure and a clear path to access, researchers can run bigger experiments, shorten iteration cycles, and collaborate more effectively across institutions.
The Digital Research Alliance of Canada is set to deliver this capacity at speed under its rapid deployment initiative, while SciNet's track record at U of T provides experienced operations and support.
Relevant resources
Digital Research Alliance of Canada
SciNet at the University of Toronto
If you're planning team upskilling alongside new compute access, explore role-based programs here: Complete AI Training - Courses by Job.
Your membership also unlocks:
