Announcing the Landmark SFU-Queen's Supercomputing Partnership

In a pivotal move for Canadian higher education and technological sovereignty, Simon Fraser University (SFU) and Queen's University have signed a memorandum of understanding (MOU) on March 26, 2026, to collaborate on designing and building a national high-performance supercomputing system. This partnership positions two leading Canadian universities at the forefront of the nation's push to develop secure, made-in-Canada artificial intelligence (AI) infrastructure. By combining their complementary strengths, SFU and Queen's aim to create a distributed network of supercomputers that will serve researchers, industry, and government across the country, ensuring that sensitive data and intellectual property remain under Canadian control.

The initiative aligns seamlessly with the Government of Canada's Sovereign AI Compute strategy, which emphasizes building state-of-the-art public infrastructure while mobilizing private sector investment. As AI demands skyrocket globally, this collaboration addresses a critical gap in Canada's computing capacity, enabling breakthroughs in fields from healthcare to climate modeling.

🚀 Canada's Urgent Need for Sovereign Supercomputing

Canada's current supercomputing landscape lags behind global leaders. SFU's Fir supercomputer, which replaced the Cedar system in 2025, holds the #78 spot on the TOP500 list of the world's fastest supercomputers—the only Canadian entry in the top 100. With over 165,888 CPU cores and 640 NVIDIA H100 GPUs, Fir supports more than 24,000 researchers nationwide through the Digital Research Alliance of Canada. However, escalating AI training requirements, which can consume exaflops of compute power, demand far more capacity.

The federal AI Sovereign Compute Infrastructure Program, set to launch in 2026 with up to $705 million in funding (part of a broader $900 million bid including private contributions), seeks proposals for large-scale systems. SFU and Queen's joint application targets this opportunity, proposing a top-10 global system at Queen's in Kingston, Ontario, and a top-25 system at SFU in British Columbia. This distributed model ensures coast-to-coast access, reducing latency for West Coast users while leveraging Ontario's energy grid.

SFU's Proven Track Record in Sustainable Supercomputing

Simon Fraser University has long been a pioneer in high-performance computing (HPC). Located on Burnaby Mountain, the Cedar Supercomputing Centre—now home to Fir—is powered entirely by clean hydroelectric energy, achieving a power usage effectiveness (PUE) of 1.07, well below the industry average of 1.56. This sustainability focus stems from partnerships with companies like Hypertec, Cerio, Corix, and Moment Energy, which enhance cooling and energy recovery.

SFU's leadership extends to innovation rankings: it has topped Canada's World University Rankings for Innovation (WURI) for five consecutive years, excelling in AI, quantum technologies, and climate research. Over 100 faculty members specialize in AI applications, from drug discovery to sustainable manufacturing. The new top-25 supercomputer will expand Fir's capabilities, potentially incorporating quantum chips by 2027, as hinted by SFU's Vice-President of Research Dugan O'Neil.

Queen's Global Expertise in Supercomputer Design

Queen's University brings unparalleled design prowess to the table. Researchers like Ryan Grant, head of the CAESAR lab—Canada's largest group on exascale systems—and Ian Karlin, formerly of NVIDIA, have contributed to top-10 supercomputers worldwide in the US, Europe, and Asia. The Centre for Advanced Computing at Queen's already serves as a key analytics hub.

Partnering with Bell Canada, Queen's will host the top-10 leadership-class system in Kingston. Bell's infrastructure support ensures scalability, with plans for energy-efficient operations mirroring SFU's model. Vice-Principal Nancy Ross emphasized how this union of HPC design and operational expertise will train future Canadian talent and drive economic competitiveness. For more details, see the Queen's Gazette announcement.

Technical Vision: A Distributed Made-in-Canada Network

The proposed systems represent a leap forward:

• Queen's Top-10 System: Hosted in Kingston with Bell, focused on exascale performance for massive AI simulations.
• SFU Top-25 System: Expanding Fir on the West Coast, optimized for regional access and quantum integration.

Canadian vendors will supply hardware, fostering domestic supply chains. The network will offer petabytes of storage and GPU-intensive resources, accessible via national allocations. As O'Neil noted, "This sovereign platform ensures Canadians can trust their data stays in Canada."

Sustainability: Powering the Future Responsibly

Energy demands of supercomputers are immense—Fir alone rivals small cities' power use. Yet both universities prioritize green tech. SFU's PUE exemplifies efficiency, recovering waste heat for campus heating. Queen's CAESAR lab specializes in energy-optimized architectures, targeting liquid cooling and renewable integration.

This aligns with Canada's net-zero goals, using BC hydro and Ontario nuclear/hydro blends. Partnerships like SFU's with Moment Energy for heat reuse set precedents, potentially reducing operational costs by 20-30% over traditional systems.

Transforming Research in Key Sectors

Access to top-tier compute will revolutionize Canadian academia:

• Healthcare: Personalized medicine via genomic simulations (e.g., protein folding like AlphaFold).
• Clean Energy: Modeling fusion reactors or battery materials.
• Defence & Public Safety: Secure AI for threat detection.
• Climate: High-res weather forecasting and carbon capture optimization.

James Peltier, SFU's Director of Research Computing, highlights applications in green tech to combat climate change. Industry users, paying fees via academic portals, will co-develop solutions, bridging lab-to-market gaps.

Explore SFU's current impacts in their news release.

Economic Ripple Effects and Job Creation

This $900 million investment promises thousands of high-tech jobs: HPC engineers, AI specialists, data scientists. Universities will train via new programs, with SFU's 100+ AI faculty expanding cohorts. BIV reports economic multipliers from industry access, akin to US national labs.

By retaining IP domestically, Canada avoids "compute colonialism," where foreign clouds dominate. Queen's system, backed by Bell, could spawn spinouts in Kingston's tech corridor.

Cultivating AI Talent in Canadian Higher Education

Higher education stands to gain immensely. Shared access democratizes compute, enabling smaller universities to participate. Programs like SFU's AI graduate streams and Queen's exascale courses will scale, producing experts for a market projected to need 20,000 AI pros annually by 2030.

International collaborations, per Digital Research Alliance, will attract global talent while prioritizing sovereignty. Check the federal strategy for broader context.

Challenges Ahead and Strategic Solutions

Power supply, skilled labor shortages, and procurement timelines pose hurdles. Solutions include:

• Phased builds starting 2027.
• Private funding to supplement federal dollars.
• Workforce pipelines via university-industry apprenticeships.

Competition from US/EU bids is fierce, but Canada's clean energy edge differentiates it.

Photo by K. Mitch Hodge on Unsplash

Future Outlook: Positioning Canada as an AI Powerhouse

If funded, operational systems by 2028 could vault Canada into TOP10 nations for HPC. This bolsters higher ed's role in national security and innovation, inspiring similar inter-university pacts. As Ross stated, it ensures Canada thrives in an AI-driven world.

For researchers eyeing opportunities, this heralds a new era of computational abundance in Canadian academia.