Hello,

This week, Canada’s push for more AI computing power is running into a local question: who bears the costs when a data centre moves in?

Plus: Quebec AI financing, deepfake investment scams and new Canadian guidance for AI agents.

Here’s the week.

THE BIG ONE

Canada wants more AI infrastructure. Many Canadians don't want it next door.

What happened

Canada currently has about 337 megawatts of AI data-centre capacity. A federal investor presentation identified more than 20 gigawatts of projects under planning or development, but Ottawa cautioned that the figure was only a snapshot of proposals at very different stages—and that most of the proposed capacity is not expected to proceed.

Opposition is growing alongside the buildout. Residents have raised concerns in communities from Vancouver to Olds, Alberta, and Hamilton. In Manitoba, the province rejected a proposed hyperscale facility near Île-des-Chênes after concerns about environmental effects, electricity use, local impacts, and limited long-term economic benefits.

Why it matters

Canada needs far more computing capacity if domestic companies and institutions are to rely less on foreign infrastructure. But while the strategic benefits may be national, many of the costs—electricity demand, water use, land, noise and grid upgrades—are concentrated in host communities.

In an Angus Reid Institute survey, 68% of Canadians said they would oppose a large AI data centre within a few blocks of their home.

Your move

If a data centre is proposed near you, look beyond the investment total. Check municipal planning agendas and ask five practical questions: Where will the electricity come from? How much water will it use? Who pays for grid upgrades? How many permanent jobs will remain after construction? And what benefits are guaranteed for the host community?

Our read

Canada does need more computing capacity. But "sovereign AI" will be difficult to defend if communities absorb the infrastructure costs while ownership, control, and most of the economic benefits remain elsewhere. Transparent local-benefit agreements should be treated as part of the project — not as an afterthought.

THE WEEK IN BRIEF

Ottawa is backing 63 Quebec AI projects

Canada Economic Development for Quebec Regions announced $13.85 million in financial contributions for 63 AI projects across the province. Sixty of the contributions are repayable and three are non-repayable, under the Regional Artificial Intelligence Initiative.

Why it matters: This shows what AI-adoption policy looks like when it reaches individual organizations—but it is also a reminder that government support does not always mean grant money. Most recipients will be expected to repay the financing.

AI impersonations of the Prime Minister are being used in investment scams

Canadian regulators have warned about fraudulent articles and social-media posts claiming that Prime Minister Mark Carney endorses unregistered investment platforms. The Canadian Anti-Fraud Centre says Canadians reported more than $704 million in fraud losses in 2025, including $351 million connected to investment fraud.

Why it matters: A convincing video, article or endorsement is no longer proof that a public figure supports an investment. Verify the platform’s registration independently before sending money or personal information.

OSFI outlines safeguards for generative and agentic AI

Canada’s federal financial regulator has published sound practices for banks, insurers and other regulated institutions using generative and agentic AI. They include limiting what AI agents can do autonomously, restricting sensitive information in public tools, monitoring outputs, creating approval checkpoints and keeping humans accountable for important decisions.

Why it matters: A chatbot produces answers; an AI agent can use software and take actions. When an error can affect a payment, customer record or financial decision, permissions, monitoring and human approval become as important as the model itself.

IN PLAIN LANGUAGE

Why do AI data centres use so much electricity?

AI systems run on large numbers of specialized computer chips working at the same time. Training a model can require enormous clusters of those chips, while serving millions of requests keeps the systems operating around the clock. The chips also generate heat, so data centres need cooling, networking equipment, backup power and extra capacity for reliability.

One useful distinction: a project’s announced power capacity is the maximum it is designed to draw—not necessarily what it will consume every minute. Even so, scale matters. AI may feel like software, but it depends on physical infrastructure: buildings, power lines, cooling systems, water infrastructure and land.

ONE THING TO TRY THIS WEEK

Run a three-minute investment-scam check

When an ad shows a politician, celebrity, bank, or news outlet promoting an investment, don't use the link inside the ad. Search for the company independently, then check whether the person or firm is registered through the Canadian Securities Administrators.

Treat guaranteed returns, urgent deadlines, and requests to move the conversation to a private messaging app as warning signs. A convincing video can be generated. Registration and independent verification are much harder to fake.

Get the Weekly Brief

A calm, practical briefing on the AI developments that matter in Canada—delivered every Tuesday.

WORTH READING

  1. The Canadian Press — Reporting on the scale of Canada’s planned data-centre expansion and the local opposition forming around it.

  2. Angus Reid Institute — Survey findings on AI regulation, employment and having a large data centre nearby.

  3. Government of Canada — The official list of 63 Quebec AI projects and whether each contribution is repayable or non-repayable.

  4. OSFI — Safeguards involving autonomy, oversight and sensitive information.

  5. Canadian Securities Administrators — Guidance on recognizing deceptive investment promotions and verifying registration.

CONTINUE READING

  1. What is AI literacy, and why does Canada suddenly care?
    A plain-language guide to understanding what AI tools can do, where they fail and when human judgment is essential.

  2. AI for All, explained
    What Canada’s national AI strategy says about workplace adoption, skills, responsible use and public trust.

  3. The Weekly Brief — Issue 02
    Four Canadian AI developments worth knowing this week.

That's the week. See you next Tuesday.

— Padge T.

AI Brief Canada is an editorially independent publication published by Groundshift Advisory Inc. and written by Padge T.

Groundshift Advisory helps Canadian organizations adopt AI responsibly. Sponsored, affiliated, or commercially connected content will always be clearly disclosed.

Issue disclosure: Nothing in this issue is sponsored, affiliated, paid, or commercially connected.

AI use: AI tools supported research, source discovery, outlining and drafting for this issue. All material was reviewed, edited and fact-checked by Padge T. before publication. Read our AI Use Disclosure.

Not affiliated with the Government of Canada.

Corrections or feedback: We welcome corrections and reader feedback. Email [email protected] or review our Editorial Policy.

Keep Reading