The federal Liberals are attempting to address Canadian concerns over the impact of large-scale artificial intelligence data centres by introducing a set of voluntary industry-backed “principles” that they say will guide new facilities’ construction.
At the same time, the government is acknowledging that “we don’t have the jurisdiction” to enforce those principles.
In Markham, Ont., Thursday, AI Minister Evan Solomon unveiled the federal government’s five-point “national framework” for how data centres should be built across the country. The move comes after high-profile opposition to new AI-servicing facilities in multiple Canadian cities.
The set of principles are entirely voluntary, backed by some of the largest international companies pursuing generative artificial intelligence programs, and are not binding in any way. Solomon stressed that decisions on data centre approvals and locations will still be up to provinces and municipalities.
The proposals suggest the federal government would prefer data centres create “lasting local benefits,” not shift electricity costs onto Canadians, minimize water use and “environmental impacts,” be transparent about their “local impacts” and bring strategic value to the country.
“All the questions Canadians are asking about how are we going to build the digital infrastructure of the future are excellent questions … we get them,” Solomon told reporters at a press conference.
“What will AI do for my community? Who is it going to benefit? Where is it going to be built? How much water will it use? Is it going to affect my power bill? Is there going to be a net benefit to my community?”
The federal framework has the endorsement of the Federation of Canadian Municipalities, and “signatories” include U.S. web giants Amazon, Anthropic, Google, and Meta, as well as Bell Canada and TELUS.
Asked if the federal government would monitor whether companies are complying with the framework, Solomon suggested they would not.

Get breaking National news
Get breaking Canada news delivered to your inbox as it happens so you won’t miss a trending story.
“Any data centre first requires power allocation, that’s the provincial regulator, then they’re going to go to a municipality and say ‘we’d like to locate here,’” Solomon said.
“And what we are saying in the (federal) framework … is to provide reliable and verifiable information on the standards that they set. Those are going to be enforced by the municipalities or the provinces, because that is their enforcement, we don’t have the jurisdiction.”
A July study released by York University researchers Lyndsey Rolheiser and Alexander Carlo found that while there are just four active AI-specific data centres operating in Canada, there are an additional 49 facilities either announced or currently under construction.
The rapid proliferation of AI data centres — required by U.S. companies locked into an artificial intelligence arms race against their rivals — has led to significant opposition in America.
Community concerns generally revolve around the massive water and electricity consumption used to run the facilities, as well as noise pollution inflicted on nearby homes from backup diesel generators, constantly running electronics, and cooling fans.
The massive amount of power used by AI data centres have also put additional stress onto aging North American energy infrastructure, increasing the risk of blackouts.
AI security intelligence company 10a Labs tracks local opposition to data centre construction. In the company’s most recent Data Centre Watch Report, they estimate that at least 75 projects marketed as worth $130 billion U.S. were “disrupted” by “local opposition” to the projects in the first quarter of 2026 alone.
Canada hasn’t seen that level of widespread opposition, largely because the country hasn’t seen the same scale of proposed data centre projects to date.
Opposition to new data centre construction in Hamilton, Ont., became a flashpoint earlier this year — but that city’s council voted in July to reject a one-year moratorium on development.
There are signs, however, that Canadians are souring on the prospect of AI products in their lives.
Ipsos’ annual AI Monitor report, released in June, found that Canadians are among the least enthusiastic populations for the technology out of 32 countries surveyed.
The reported noted that the share of Canadians who expressed excitement about AI products and services was just 26 per cent, down five percentage points from the year before. Canadians are among the most nervous when it comes to the technology — with 67 per cent saying so, a full 17 percentage points higher than the international average.
Few Canadians believe AI will improve their country’s economy (14 per cent) or the job market (11 per cent) in the medium term, and just 27 per cent trust companies that use AI will protect their personal data.
All that makes for an uphill battle for Solomon, tasked by Prime Minister Mark Carney to pitch the broader adoption of AI tools across the public and private sector.
“There’s a big rush to build because of this new technology, and you gotta build right,” Solomon said.
“We want to attract the investment, but be transparent about it. Though we need to build at speeds we haven’t seen before, (that’s) not at the expense of our values.”
As Solomon was making his announcement Thursday, the website Downdetector reported outages affecting OpenAI’s ChatGPT, Anthropic’s Claude, and X.com’s Grok artificial intelligence services.
With files from Global News’ Sean Boynton.
© 2026 Global News, a division of Corus Entertainment Inc.


