64% of respondents, given a direct choice between two positions, said their provincial government should limit the development of large data centres. A national survey released on October 7, 2026 put that number on the argument that has followed every large data centre proposal this year. Only 18% said their government should encourage the development of more of them. The release of that figure changes no permit, pauses no construction and answers no application. It does something quieter and harder to reverse: it fixes the political weather in which every future application will be heard.
The survey put the question to 4,417 Canadian adults online between September 8 and September 16, 2026, and its figures hold to within 1.5 points. The question first told respondents what counted: a facility built or proposed to carry cloud computing, artificial intelligence and other digital services. That wording matters more than the headline it produced, because it defines what the majority actually agreed to.
The short answer. On the question as asked, Canadians side with limits by a wide margin: 64% for provincial limits on the development of large data centres, against 18% for encouraging more. The majority holds in every region surveyed, from 72% in Manitoba and the Atlantic provinces down to 50% in Quebec. The figure states a preference about provincial policy toward large builds. No ban was offered in the question, no project was named in it, and no local approval process shortens because the number exists.
A number this size decides nothing on its own. Every participant at the next hearing will argue as if it decides everything.
- What exactly did the question ask?
- Where does the wish for limits run strongest?
- What sits underneath the number?
- Which public is telling the truth?
- What does the 64% change at the next hearing?
What exactly did the question ask?
The wording builds three fences around the result, and honest readers stay inside them. The first fence is scale: the question concerned large data centres and said nothing about smaller facilities. The second is jurisdiction: it assigned the decision to provincial governments and to nobody else. The third is the choice itself, which ran between limiting development and encouraging more of it, the two positions taking 82% of respondents between them. Nobody was asked about a building proposed for their own street, and nobody was offered a prohibition to vote for. Executives who quote the 64% as a public rejection of data centres read well past the question, and executives who dismiss it as noise read well past its margin.
Where does the wish for limits run strongest?
Manitoba and the Atlantic provinces lead the country at 72% each. British Columbia and Saskatchewan follow at 70% each, Ontario at 69%, and Alberta, the province carrying the heaviest current construction argument, returns the national figure at 64%. Alberta’s result rests on a provincial subsample of 415 respondents inside the national sample of 4,417 adults. Quebec closes the table at 50%, the only region where support stops at exactly half. A premier reading this table finds no regional shelter in it, and a proponent reading it finds no friendly province waiting.
What sits underneath the number?
Among respondents who favoured limits, the leading reason named was high water use, at 51% of that group. The detail deserves more weight than it usually gets, because it describes the shape of the objection itself: the recorded objection is a resource objection, stated in terms of a measurable input that every project consumes. Resource questions can be answered with project-level facts, published in advance and tested in a public room. What the survey records is a public asking, in majority numbers, who carries the resource bill for the next round of building.
Which public is telling the truth?
One number, 64%, is currently carrying two opposite arguments, and each side can point to real evidence. The first holds that the public has delivered its verdict against the builds: 64% want limits, and councils have begun pausing applications, as Mississauga did in September. The second holds that the same public keeps deepening its reliance on everything these buildings carry, from cloud storage to AI services threaded through ordinary office work. Construction already approved continues across the country, and demand for the underlying capacity has not paused. Both beliefs describe the same country accurately, because they describe two different acts that share one careless name. The survey measured a preference about provincial policy, stated in the abstract, before any particular building had an address. No real project is ever national in that way, because every build arrives one town at a time, carrying its own water demand, its own grid connection and its own council vote. Consent for data centres in Canada is a condition each project earns town by town, on facts proved in each successive room, and no single national reading settles it for the whole country. The 64% sets the terms on which that consent gets negotiated, starting with the next application filed. (Mississauga Had One Data Centre Application. It Froze Every Lot in the City.)
What does the 64% change at the next hearing?
From October 7 forward, every proponent carries this number into the hearing room. Councils will read it back to applicants, opponents will open their remarks with it, and Ottawa’s five September expectations now have a measured public mood standing behind them. The practical answer is evidence rather than messaging, assembled before the room asks for it. Water figures, power figures and a local benefit case, stated in writing and specific to the project, are the only currency this number accepts. Consent work belongs inside the build schedule, named and dated like any other deliverable, because the council across the table has read this release and will ask who owns it. Ottawa Set Five Expectations for Data Centres. The Switches Sit Elsewhere.
None of the coming decisions move on a pollster’s calendar, or on Ottawa’s. They land where the last ones landed, in council chambers, in provincial connection queues and in regulators’ dockets, on dates set locally by people who can be voted out of the room. The survey will still be quoted in those rooms a year from now. The projects standing then will be the ones whose facts were ready before the quoting started.
Recap. A survey released October 7, 2026 found 64% of Canadian adults want their provincial government to limit large data centres, against 18% who would encourage more. High water use leads the reasons named by those who favour limits. The figure settles the national mood and decides no individual project. Consent is earned in each successive room, on each project’s own facts.
Read the record this number now stands over: Canada’s data centre map: what is built, what is blocked, what is waiting.
