MaRS Discovery District:How North America's largest urban innovation hub became the commercialization layer between Canada's AI research and Canada's AI companies
What they're doing
MaRS Discovery District — North America's largest urban innovation hub — has moved from generalist startup support to running AI-specific programs anchored in its Toronto campus. Supply AI at MaRS supports cohorts of up to 12 Canadian AI ventures over six months with advisory services, capital access, market intelligence, and dedicated AI/ML consulting. The inaugural cohort included companies applying AI across supply chain operations: Basetwo, Mercator AI, Mely.ai, and nine other ventures — a deliberate bet on sector-depth over breadth.
In June 2025, MaRS co-hosted Frontiers of AI alongside the University of Toronto and the Vector Institute — a full-day event connecting researchers with practitioners. The three-institution collaboration reflects MaRS's role as the commercialization bridge in Toronto's AI cluster: Vector builds research capacity, U of T produces talent, and MaRS turns both into companies. Together they anchor a pipeline that's made Toronto one of a small number of cities globally that can claim a full AI value chain from foundational research to venture scale.
In May 2026, Canada's Minister of Artificial Intelligence and Digital Innovation visited MaRS as part of the federal government's push to accelerate AI innovation across the Greater Toronto Area. MaRS is also a Scale AI partner organization, connecting portfolio companies to Canada's AI supply chain supercluster and the broader federal AI strategy.
Why it matters
Toronto's global AI profile didn't emerge from a single investment or policy decision — it emerged from 15-plus years of institution-building across MaRS, Vector, and the University of Toronto. MaRS's specific function is commercialization: taking research-grade AI capability and building the advisory, capital, and market-access infrastructure to turn it into Canadian companies. That function — and whether Canada retains the companies it builds — is what long-term Canadian AI competitiveness actually hinges on.
What's working
- Tri-institution coordination with Vector and U of T means the pipeline from research to venture to scale has institutional support at every stage, not just at one point
- Sector-specific programs like Supply AI create genuine peer cohorts where companies are solving adjacent problems — the learning density is higher than in generalist accelerators
- Ministerial-level government engagement in May 2026 confirms MaRS has become part of Canada's official AI infrastructure, not just a Toronto landlord for startups
What's harder than it looks
- Canada retains a fraction of the AI companies it grows — many eventually raise US capital and shift headquarters south; MaRS can support companies, but can't by itself solve the scale-capital gap that pulls growth-stage ventures toward US VCs
- Balancing early-stage founders who need hand-holding with later-stage companies who need market access across a large and diverse portfolio is genuinely difficult to get right simultaneously
- Measuring the counterfactual impact of ecosystem infrastructure — what would have happened without MaRS — is effectively impossible, which makes it hard to defend continued public investment with the same precision other programs can
Three takeaways for similar orgs
- 01Ecosystem infrastructure is AI adoption infrastructure: the innovation hubs, shared labs, and commercialization programs that surround a technology cluster determine whether that technology stays and compounds locally or migrates to wherever capital is cheapest
- 02Sector-specific cohorts beat generalist programs for AI because the technology choices, data challenges, and customer acquisition paths differ enough that peer learning only works within a narrow enough domain
- 03The Toronto AI hub model — anchor university plus research institute plus commercialization hub — is a template other Canadian cities are now trying to replicate; MaRS's operating decisions carry lessons for Calgary, Vancouver, and Montreal