Why are many new data centres being planned in Canada? Your questions answered

Also: are there health risks from living near a data centre?

This idea came from listeners like you who contacted us. Send us all your questions about data centres and the environment. We are listening at ask@cbc.ca.

In 2026, AI feels everywhere, and some Canadians wonder if they want huge AI buildings humming in their backyards.

Canada has bet heavily on AI, leading to many new AI data centre plans that need lots of water and power.

Supporters say the sites could keep Canada a leader in AI. Yet fast growth makes some towns pause. In Ontario, Oakville voted in August to stop data centre building for a year, and Mississauga will vote in September on a similar pause. Councillors acted after a proposal for a data centre the size of three football fields.

On Thursday and Friday, hundreds attended town meetings to ask Alberta’s tech minister about the province’s AI data centre plans.

Many CBC readers have also asked why so many data centres are suddenly being built in Canada.

Here’s what we know.

A data centre is a building that holds and powers computer servers and networking gear needed for computing and storing data. Whenever we visit a website or stream a movie, a physical server handles that request for us.

AI data centres in the news are far larger and use more power than normal data centres. They run powerful chips called GPUs, which raise energy use and need stronger cooling to handle the heat they create.

Canada wants its own AI data centres. What does that really mean?

What’s the difference between normal and AI data centres? A professor from Waterloo says it’s the computers they use.

What is causing the growing opposition to AI data centres?

“For a traditional cloud building, we might have a few megawatts of power, but AI sites often need hundreds of megawatts,” said Anne Pasek, Canada Research Chair at Trent University, speaking to CBC News.

These sites can cover thousands of square feet, and “hyperscale” data centres are becoming common. Some define hyperscale as 50 megawatts, others as 100. Just 50 megawatts can power 37,500 homes.

“The AI facilities we are building keep getting bigger, both in land area and in energy and water use,” Pasek explained.

Demand for AI has shot up, and Canada wants to be a leader, hoping to draw investors and boost the economy.

In 2024, up to $700 million was set aside to support new or expanded data centres.

Prime Minister Mark Carney launched Canada’s national AI plan in June, called “AI for All,” promising at least $2 billion in new funding. One goal is to raise AI use in Canadian firms from 12 % now to 60 % by 2034.

Carney’s AI plan talks a lot about jobs and training but gives few details.

Supporters say AI growth is inevitable and Canada is better off leading than falling behind.

“It’s part of modern life,” said Martin Karsten, a computer science professor at Waterloo, to CBC News Kitchener‑Waterloo on Aug. 18. “If people keep wanting AI services, we have to accept that these data centres will keep arriving.”

Many towns push back, worried about water, power and health. Critics also ask if data centres bring enough local economic benefit, noting they create short‑term construction jobs but little lasting industry.

Some experts warn the rush to fund AI could be a bubble that bursts, leaving AI data centres as stranded assets. Some AI‑focused companies are now pulling back because costs rise, raising doubts about the technology’s return on investment.

Canada now has only five hyperscale data centres of at least 50 MW, but a York University study says 159 sites were proposed or under construction by July. Of those, 116 are expected to exceed 100 MW – a first for Canada.

AI agents are far behind human workers. Why are tech firms laying off people?

Their goal was to use as much AI as possible. Now some firms are pulling back.

The study shows the industry is growing fast. Over 200 new projects were announced, with an average size of 122 MW – more than ten times the average size of Canada’s 194 active sites.

Most active data centres are in Ontario, then Quebec and British Columbia, but most new proposals are in Alberta. If all are built, Alberta’s share of Canada’s total data‑centre capacity would jump from 3 % to 93 %.

“If you look at where the data centres sit, the electricity a new centre would use is similar to the total use of the nearby community,” said Alexander Carlo, a post‑doc at York University’s Schulich School of Business and co‑author of the study, to CBC News.

Huge AI data centres are coming to Canada. One province sits at the centre of this growth.

Data centres made up 4.4 % of U.S. electricity use in 2023, according to the U.S. Department of Energy, which forecasts this could rise to as much as 12 % by 2028.

During a heat wave or snowstorm, power grids may struggle with the extra load from a data centre, Pasek said, raising the chance of brownouts. The extra strain could also push up consumer electricity bills.

A hyperscale AI data centre boom is on its way.

Computing gear creates a lot of heat, so facilities use water to cool it. Many cooling systems evaporate that water, sending it into the atmosphere instead of returning it to the local water cycle.

“They can take water from a watershed, let it evaporate, form clouds, and those clouds may rain far away, not back into the original watershed,” Pasek explained.

Small sites may pull tens of thousands of litres a day, while hyperscale sites can draw over a million litres daily.

A 2025 report by the U.K. Government Digital Sustainability Alliance found a single 100‑MW hyperscale site could use about 2.5 billion litres of water each year – enough for 80,000 people.

AI data centres use huge water volumes, but measuring exactly is difficult.

Some regions can handle that demand, but drought‑prone areas could face problems.

“Many places are already water‑poor. Adding even a small extra demand can have big impacts,” said Kaveh Madani, director of the United Nations University Institute for Water, Environment and Health, to CBC News in August.

Why people near data centres worry about water.

Ontario Premier Doug Ford said in August that the province’s data‑centre plan would favor closed‑loop cooling systems that recycle water and cut intake compared with evaporative cooling.

Experts say risks exist, but they differ by size, design and power source, so different data centres carry very different risks.

Air pollution

Data centres can pull power from local grids, but connecting them fully can take years. Many are now built with on‑site power plants that run on natural or methane gas. When powered by fossil fuels, data centres emit gases that add to climate change and release fine particles called PM2.5.

“We know that breathing fine particles raises the risk of many health problems, including lung, heart and brain issues,” said Francesca Dominici, a Harvard biostatistics professor, to CBC News. She added that long‑term exposure can increase hospital visits and worsen asthma.

An independent study she co‑authored with EmPower Analytics looked at an AI data centre near a residential area in northern Virginia and found its eight natural‑gas turbines and 51 diesel generators could cause $53‑$99 million US a year in health‑related damages.

Research suggests long‑term exposure to PM2.5 may also raise the risk of lung cancer.

Noise pollution

Like other industrial sites, data centres make constant noise from fans and backup diesel generators. They can also emit a low‑frequency hum, and long‑term exposure can cause stress, sleep issues and other health problems, Dominici said.

NBC News reported that residents have filed at least 15 lawsuits in 12 U.S. states over constant data‑centre noise in the past three years.

Why Toronto councillors and residents are pushing back against data centres

Concerns have also been raised that closed‑loop cooling could affect water quality if wastewater is mismanaged. Much is still unknown about overall risks, Dominici said, which is why she thinks development should pause for more research.

“There has been a race to build these AI hyperscalers without really thinking about possible negative health impacts on nearby communities,” Dominici said.

Experts say it’s a trade‑off question. One option is to make data centres more sustainable.

Health risks from air pollution, for example, can be reduced if a data centre and its backup power use wind or solar instead of fossil fuels.

“If you could make the facilities as efficient as possible and power them with renewable energy, either directly or stored in batteries, we’d make real progress,” said Daniel Bresette, president of the U.S. Environmental and Energy Study Institute, to CBC News.

But that option costs more, so developers may lack incentive without rules or community pressure. Most new Alberta projects plan to use natural gas, including a $13‑billion AI data centre announced by Meta last month.

Reducing noise through design also adds cost. Some U.S. regions are adding zoning rules based on noise and require developers to do noise studies before building.

Placing AI data centres in remote areas can lower risk to the general public, but it may shift impacts onto rural and Indigenous communities.

A proposed $70‑billion gas‑powered data centre in northern Alberta is being challenged by the nearby Sturgeon Lake Cree Nation, which says it was not properly consulted about a water licence.

UN report highlights environmental impacts of AI data centres

Another option is to spread AI computing capacity more widely instead of concentrating it, Pasek said, to “break up the burden” on communities.

The third option is to “rethink fundamentally what we want AI to be,” she said. This could mean returning to smaller, open‑source projects rather than huge AI systems that only profit when pushed onto businesses and Canadians.

COPYRIGHT DISCLAIMERS

4English does not own the rights to this content. All rights belong to the owner. 4English only uses this content to help users learn English via the app's features. Contact to request the removal of your content: contact@4englishapp.com.

Prática de escuta

Prática de escuta