Canada Throws $500M at AI — and Signs Britain to Watch the Risks

Published October 13, 2026 · By Daniel Rao

Canada had quite the day yesterday. Ottawa pushed out two major AI announcements, both dated October 13 — one from London, Ontario and one from London, England — and if that coincidence made you do a double-take, you weren't the only one. The government spent the day writing cheques and writing rules: another $500 million over five years to get Canadian businesses actually using AI, plus a renewed AI-security pact with Britain covering frontier-model risks, evaluations and technical standards.

Here's the bit that makes the money interesting. Canada didn't spend $500 million because it's winning at AI. It spent $500 million because, by its own government's admission, it's not. Barely 12% of Canadian companies were using AI in mid-2025. KPMG's international research ranks Canada 44th out of 47 countries on AI literacy and training — forty-fourth. The cheque is an admission ticket: Ottawa knows exactly how far behind it is, and it's paying up to close the gap.

And that's what makes yesterday's double announcement worth your time. Most governments do one of two things: subsidize the industry or regulate it. Canada tried to do both on the same day. The money says "AI is infrastructure." The Britain deal says "AI is a frontier risk." Both can be true. But only one of them comes with a five-year budget.

The Cheque: $500M for "AI for All"

The headline number: an additional $500 million over five years to expand the Regional Artificial Intelligence Initiative (RAII), bringing total federal funding for the national initiative to $700 million. The announcement came from Evan Solomon, the minister of Artificial Intelligence and Digital Innovation — and, in his other hat, the minister responsible for FedDev Ontario, the regional development agency that delivered the news in London, Ontario.

The money will flow through Canada's regional development agencies, and the priority sectors read like a greatest-hits of the Canadian economy: health and life sciences, energy and natural resources, transportation, agriculture, manufacturing and robotics, plus regional specialties. The stated aims are the ones you'd expect — help Canadian AI businesses find customers, attract investment, commercialize and scale Canadian-made products — plus one less expected: broader access for Indigenous Peoples, youth and underrepresented groups.

Context matters here. Southern Ontario — where the announcement landed — is home to more than half of Canada's AI firms, which tells you this wasn't really a regional story dressed up in regional clothing. This is the national strategy, AI for All, with a postmark. The strategy, announced in June 2026, promises to build trust, strengthen Canadian AI capabilities and sovereignty, and make sure Canadians can use AI responsibly — with the funding flowing through the LIFT program, the Compute Access Fund and the expanded RAII.

Solomon's framing was blunt. "The question is not whether AI will reshape our economy," he said in the release. "It is whether Canada will capture the opportunity." The rest of the quote is the standard pitch — RAII will help businesses adopt AI and let Canadian AI firms scale from home, keeping technology, talent and economic value anchored in Canada and under Canadian control.

A $1M Bet on Lung Ultrasounds

Every funding announcement needs a prop, and yesterday's had a good one. At the London event, parliamentary secretary Peter Fragiskatos announced a FedDev Ontario investment in Deep Breathe Inc., a London startup founded by Dr. Rob Arntfield that makes an AI-powered lung ultrasound system — pitched as a faster, cheaper, portable alternative to chest X-rays, with less radiation to boot.

The shape of the deal: a project worth more than $2 million, with FedDev Ontario putting in over $1 million, to scale the system. It's a small cheque next to the $500 million headline, but it's the honest version of what the big number is supposed to do — take a Canadian AI product, built by a Canadian company, and push it toward actual patients and actual revenue. The subtext is simpler: Canada doesn't want to export raw AI talent to California anymore; it wants the companies to scale at home.

Applications for the expanded southern Ontario funding run until January 19, 2027, with projects that have eligible costs before March 31, 2027 needing to be in by November 24, 2026. The bureaucracy, at least, is moving at startup speed.

Photo: Parliament Hill in Ottawa. Wikimedia Commons, CC BY-SA 3.0.

Atlantic Canada Gets Its Cut

The $500 million is national, but yesterday also had a regional twin. In Moncton, New Brunswick, Justice Minister Sean Fraser — in his capacity as the minister responsible for the Atlantic Canada Opportunities Agency (ACOA) — announced nearly $40 million more for Atlantic Canada AI adoption and commercialization, bringing ACOA's total RAII delivery to over $55 million.

The mechanism is an Atlantic-wide call for proposals aimed at small and mid-sized firms with AI projects valued at $1 million or more, judged on measurable outcomes: productivity gains, competitiveness, commercialization success, growth. It closes November 20, 2026. Fraser's quote carried the same edge as Solomon's: "AI is changing how businesses compete, and we can't afford to see Atlantic Canada fall behind."

The release cited the Atlantic Economic Panel's recent report, Bolder Ambitions: Greater Prosperity for All, which argued Atlantic businesses need help turning AI from slideware into growth. Fair enough — but it also quietly confirms the strategy's geography problem. The AI firms are in southern Ontario. The money has to be pushed east to keep the federation pulling in the same direction.

Meanwhile, in the Other London

Now the other announcement — and, if you squint, the more consequential one. Also dated October 13, but this one datelined London, England: Canada and the United Kingdom renewed their partnership on AI security, with Solomon in London for the occasion. We checked the canada.ca release — no ministerial quote mentions the UK deal, and no UK mention appears in the funding releases. The two announcements ran as parallel operations, not a coordinated campaign. Which, honestly, makes the coincidence better.

The pact recommits the Canadian AI Safety Institute and the UK AI Security Institute to cooperation, and it builds on the UK-Canada Science of AI Safety Partnership first announced in May 2024, plus the Growth and Innovation partnership signed in 2025. The substance falls into three buckets:

1. AI security standards — interoperable standards for recognizing and spreading sound AI security practices, with an eye to future international cooperation.

2. Enhanced information-sharing — including through appropriate Five Eyes channels for sensitive reporting, plus shared lessons on domestic resilience. Five Eyes intelligence-sharing on frontier AI model risk is the sentence most people will read twice.

3. Expanded technical cooperation — regular exchanges on frontier AI capabilities, emerging risks and evaluation methodologies; staff exchanges; coordinated evaluations of models.

The timing isn't accidental. Last month, PM Burnham committed to putting AI at the heart of the UK's G20 presidency next year; in September, Carney signed the Call for Control of Frontier AI Models launched at UNGA; and back in June, the Five Eyes jointly warned organizations about rapidly evolving AI-driven cyber risk. This pact slots into a busy autumn of institution-building — and it gives both countries a standing arrangement to compare notes on the models that scare them. It also follows the UK's recent AI dealings with Washington: London is building AI relationships in every direction at once.

This isn't abstract for Ottawa, either — Canada has already lived through Canada's earlier scare with AI agents breaching government systems, which is part of why a standing pact with Britain's safety machinery matters more than it looks. Rather than building everything from scratch, Canada is attaching itself to an institutional setup that has been running for years — the sensible move of a smaller player.

The Named Enemy: Dependence

What ties the cheque and the pact together is the problem Carney's government keeps naming out loud: dependence. Canada's strategy documents say the country is heavily dependent on foreign suppliers for compute, cloud and data storage — which means foreign entities potentially accessing Canadian data and tilting the field against Canadian firms. That's unusually blunt language for a government press release, and it's the honest core of AI for All: this is a sovereignty project dressed as an economic one.

Photo: Canadian Prime Minister Mark Carney at the G7 summit in Kananaskis in June 2025. Simon Dawson / No 10 Downing Street, CC BY 2.0.

The numbers Carney's team chose to quote in June are striking: more than 3,500 Canadian AI companies have raised C$37 billion-plus in venture capital; generative AI could add C$187 billion a year to the Canadian economy by 2030; roughly 12% of Canadian companies were using AI in mid-2025, against a target of 60% by 2034; and the plan promises 250,000 new jobs from AI adoption by 2031, free AI literacy training for all Canadians, and up to 90,000 AI placements for young Canadians.

Then the other numbers — the ones that sting. Only 31% of Canadian organizations have fully integrated generative AI into core workflows, per KPMG. And that 44th-of-47 AI literacy ranking. Canada is, by the government's own chosen metrics, near the bottom of the class it claims to be leading. The strategy is a confession wearing a plan's clothes.

Carney has also moved to get advice from the sharpest people available. On October 2 he launched the Prime Minister's National Council on Artificial Intelligence — 13 members including Yoshua Bengio, the deep-learning pioneer, tasked with advising on adoption, national champions and sovereign AI infrastructure. Bengio's presence is no accident: he has spent years on Yoshua Bengio's call for AI safety, and he remains one of the loudest voices demanding guardrails on frontier models. "Prosperity and sovereignty in the age of AI belong to nations that can build, adopt, and govern AI on their own terms," Carney said at the launch.

Is 60% by 2034 Believable?

Here's where a reporter is allowed to have an opinion, so here's mine: the 60%-by-2034 adoption target is the most honest number in the whole strategy, and also the least believable.

It's honest because it admits the starting point — 12% — without flinching. Most governments would have buried that number. Canada put it in a strategy document. A target drawn as a straight line from a real, measured, embarrassing baseline deserves credit.

It's not believable because the gap between 12% and 60% is a chasm, not a slope, and $500 million over five years — spread across regional development agencies, six priority sectors, literacy programmes and everything else — is a rounding error next to what American and Chinese firms spend on AI in a single quarter. Canada's 3,500 AI companies raised C$37 billion in venture capital; the federal government is now offering a fraction of that, over half a decade, to fix the adoption problem venture capital didn't.

But here's the thing: the money was never the whole bet. The UK pact, the Bengio council, the Five Eyes information-sharing — that's the part of yesterday that might actually matter in 2034. If Canada can't outspend the giants, it can out-coordinate the middle powers, and a Canada–UK axis on AI safety standards is a genuinely smart move for two countries that are too small to set the terms alone and too serious to let someone else set them.

The cheque buys time. The pact buys leverage. Only one of those compounds — and it's not the one with the dollar figure.

Sources

This article is based on canada.ca (the Regional AI Initiative expansion, the Atlantic ACOA announcement and the Canada–UK AI security partnership releases), gov.uk (the UK–Canada AI security partnership), pm.gc.ca (the National Council on Artificial Intelligence) and newswire.ca (press-release copies); with context from the Wall Street Journal's June 2026 coverage of Canada's AI for All strategy, KPMG's analysis of Canadian AI adoption and literacy rankings, refdesk.ca, and Canadian Press reporting on the National Council.

About the author: TechPulse USA's news desk covers the AI industry across the US, Europe, Canada and Australia — model launches, lab politics, and the numbers behind the hype.