Ontario municipalities are discovering they have no real playbook for the AI data centre boom landing in their backyards. As facilities proliferate across the province, cities like Toronto and Hamilton are scrambling to figure out zoning, environmental review, and grid impact after the fact rather than before construction begins.
Hamilton City Council recently voted to pause new data centre approvals, giving staff time to review the environmental and public health impacts before any more facilities move forward. Ward 3 Councillor Nrinder Nann led the push, stating in an open letter that too many negative environmental and human health impacts remain unaddressed and that stringent regulation across all three levels of government is needed before more facilities are approved in Ontario communities.
A Regulatory Gap Cities Didn't See Coming
Most AI data centres in Ontario are currently approved as-of-right under existing industrial zoning, meaning they don't require the kind of special review that their scale and power demands arguably warrant. Ontario's energy regulator, the Independent Electricity System Operator, grants grid connection approval on a first-come, first-served basis, with little coordination at the municipal level.
Toronto is now studying the issue directly. City council is reviewing whether AI data centres should require distinct zoning designations rather than falling under general industrial categories, following a motion raised earlier this year after a councillor discovered a data centre was being developed in her ward without her office being notified. Mayor Olivia Chow has agreed to sign the city onto the Global Urban Data Centres Pact, a framework aimed at setting sustainability and safety standards for the sector.
The Grid Risk Nobody's Talking About Enough
The concerns go beyond zoning paperwork. A Toronto Hydro board member and environmental lawyer warned that clustering multiple large data centres in the same area creates a genuine grid stability risk. Network equipment designed to protect data centres from power fluctuations can trip offline within milliseconds when several facilities sense the same small disturbance simultaneously, a scenario that could take down power across a much wider area than the data centres themselves.
That warning followed a real event. A fire at a downtown Toronto transmission station recently knocked out power to much of the city's core, underscoring how interconnected and fragile the grid already is before adding the additional load AI data centres bring.
Why This Matters for Business
I've advised companies for years on AI adoption, and this is the layer of the AI conversation that gets almost no attention in boardrooms. Businesses building AI-dependent products or relying on cloud infrastructure assume the physical layer, power, water, and grid capacity, is someone else's problem to solve. Increasingly, it isn't. Regulatory uncertainty at the municipal level can directly affect how quickly new data centre capacity comes online, which in turn affects compute availability and pricing for everyone downstream.
Companies with real AI infrastructure exposure, whether through direct data centre investment or through vendors who depend on it, should be tracking municipal-level regulatory movement the same way they'd track federal AI policy. The bottlenecks of the next two years are increasingly going to be local, not national.
