Prime Minister Mark Carney announced in Montréal on Oct. 29 that the federal government is taking an equity stake in Cohere, the Toronto-founded AI company, alongside private investors including NVIDIA, Bell and Schwarz Group. Cohere builds AI for business and government with an emphasis on control — where data is stored, how it is used and who can access it — and the announcement carried a commitment to at least 1,000 new high-paying Canadian technical jobs by 2030.
The phrase Ottawa used is “sovereign AI”: AI capacity that Canada builds and controls at home. Whatever you think of the policy, the phrase is about to change how AI gets bought in this country, because it hands every board and procurement team a new question to ask.
Strip away the flag and sovereignty is a data-control question. A model can be excellent and still be the wrong choice if your records sit in a jurisdiction you did not choose, are processed by people you cannot name, or can be reached by parties you never approved.
Sovereign AI is not a model you buy. It is a set of decisions about where your data lives and who can reach it — and those decisions are made during implementation, not at the keynote.
That is what makes this a deployment story rather than a funding story. Control is built during implementation: where the infrastructure sits, how the system connects to the tools already in place, what access is granted, how activity is logged and who reviews it. That is systems integration work, and it is where a sovereignty promise either becomes real or quietly evaporates.
Before your next AI purchase, ask three questions in writing: where is our data stored, where is it processed, and who can access it. If a vendor cannot answer plainly, that is your answer.