Making Sense.

Hundreds of people marched from Mount Pleasant to downtown Vancouver on October 3 to oppose AI data centres, and organizers from No AI Vancouver invited city council candidates to join them. In July, city council put off a decision on rezoning land for a proposed AI data centre in Mount Pleasant, which earlier reports say Telus would operate, and asked city staff to run more public consultation first. That consultation won't begin until after the October 17 municipal election. Four council candidates from COPE, the Vancouver Greens and Bright Future who spoke to CBC said they support a pause or moratorium on large AI data centres, citing energy use, water supply and a lack of consultation.

The debate has reached the provincial campaign too: the B.C. Greens have proposed a province-wide moratorium on AI data centres as part of their platform for the October 24 election. Because consultation won't start until after the vote, the Mount Pleasant decision is likely to fall to the next council. For businesses counting on local, Canadian-owned compute, that makes this municipal election matter more than most.

AI Minister Evan Solomon announced up to $162 million over five years for Mitacs to co-fund 10,000 paid AI work placements, pairing students, recent grads and postdocs with Canadian businesses through its ADOPT and AI+X programs, with a focus on small and medium-sized companies. The placements are part of the national AI strategy's goal of up to 90,000 AI jobs and placements for young Canadians by 2031. Businesses of any size in Canada can apply as long as AI is central to the project, and placements go to students, recent graduates and postdocs at Canadian colleges and universities.

A week later, Prime Minister Mark Carney named a 13-member National Council on Artificial Intelligence to advise on how the strategy is rolled out, including how quickly businesses adopt AI, which Canadian AI companies to back, and where sovereign AI infrastructure gets built. Those decisions will shape future funding programs like this one and where new data centres land.

Toronto-based Cohere and Waterloo-based OpenText announced at ALL IN that they will offer an AI agent product early next year to OpenText's roughly 120,000 enterprise clients, connecting Cohere's North platform and models to OpenText's data and content systems. Days later, TD committed up to $25 million over three years to a collaboration between its Layer 6 AI research centre and Cohere, focused on AI agents and better workflows.

Both deals target the same buyers: regulated industries and government, where it matters where data is stored and who controls it. Both deals build on Cohere's Canadian-built AI models. For organizations already running OpenText, a Canadian-built AI agent option may soon arrive through a vendor they already use.