Three major US technology companies control 85% of Canada's publicly available cloud infrastructure, according to a new report released ahead of the government's national AI strategy. Amazon holds 42% of the Canadian cloud market, Microsoft controls 31%, and Google accounts for 12%, according to research from the Canadian Anti-Monopoly Project (CAMP) published on June 2, 2026. This concentration far exceeds their global average of 66%, raising concerns about Canada's ability to build independent artificial intelligence capabilities.

The timing of this report is significant. The federal government is set to release its national AI strategy this week, with sovereign compute infrastructure listed as one of six core pillars. The strategy, titled "AI for All" in draft form, explicitly acknowledges that data centre and cloud options in Canada are mostly foreign-owned and will require significant investment to overcome reliance on foreign providers.

Why Does Cloud Market Concentration Matter for AI?

The dominance of these three companies, known as "hyperscalers," creates a structural dependency that goes beyond simple market competition. Hyperscalers are firms with the ability to provide computing resources on-demand globally, allowing businesses to scale from a few users to millions without investing in their own infrastructure. "The compute is available instantly, billed by the second, and distributed across a global network of data centres designed for redundancy and low latency," the CAMP report explains.

The Canadian Anti-Monopoly Project warns that this dependency carries geopolitical consequences. "Rising tensions in US-Canada relations, the intermingling of Big Tech interests with US government power, and the demonstrated willingness to use technology access as geopolitical leverage have forced Canada and others to confront an uncomfortable reality: dependence on a handful of US hyperscalers is a sovereign risk as well as a competition problem," the report states. Prime Minister Mark Carney has repeatedly emphasized this concern, noting at Davos in January that Canada is "co-operating with like-minded democracies to ensure we will not ultimately be forced to choose between hegemons and hyperscalers".

The financial stakes are substantial. Since 2021, Ottawa has spent almost $1.3 billion on cloud services provided by US companies, with most of the money going to Microsoft. Canadian small and medium-sized businesses currently have "no affordable domestic option" for computing power, forcing them to train and deploy AI models on foreign cloud platforms and send sensitive data and intellectual property outside the country.

What Is Canada's Plan to Build Sovereign AI Infrastructure?

The government's response includes multiple initiatives designed to reduce dependence on foreign hyperscalers while building domestic capacity. The draft AI strategy outlines several concrete steps:

• Canadian Tech Growth Fund: A new startup investment vehicle that will take direct government equity stakes in artificial intelligence companies, targeting smaller investments in domestic AI firms alongside the larger Canada Strong Fund, which has an initial $25 billion budget for major national champions.
• AI Compute Access Fund Expansion: A top-up to the existing $300 million program that subsidizes computing costs for Canadian small and medium-sized businesses building AI products. The government disbursed $66 million to 44 companies in the first round of payments in June 2026.
• Sovereign Cloud Computing Capacity: Prime Minister Carney has directed the federal Major Projects Office to develop sovereign cloud computing infrastructure to ensure Canadian-controlled systems for building and running AI models.
• Government as Strategic Anchor Customer: The strategy positions the federal government as a "strategic anchor customer" for domestic AI scale-ups, using government procurement as demand-side support rather than relying solely on grants.
• AI Literacy and Job Creation: The strategy commits to providing free AI literacy training to all Canadians and targets the creation of up to 90,000 AI-related jobs, with projections that AI adoption will generate up to 250,000 new jobs across the economy by 2031.

How Can Canada Reduce Switching Costs and Build Real Competition?

Experts argue that simply directing public funding to domestic companies without structural reforms could replicate existing problems. The CAMP report cautions against what it calls "maplewashed dependency," warning that "directing public funds to Canada's domestic telecommunications oligopolies without clear conditionalities for interoperability based on de facto standards would merely transfer market control to these firms".

Instead, policy analysts recommend using government purchasing power strategically. Curtis McCord, a policy analyst at CAMP, explained that the government can reshape the market without necessarily changing which companies it does business with. "And this doesn't necessarily mean that they need to change who they're doing business with, but they need to change the terms on which they do business by only buying technologies that are interoperable or substitutable," McCord said.

"What this does is it lowers the switching costs in case they eventually want to change providers. Because every provider that they work with will have a compatible system, because interoperability will solely be built in by vendors to fulfil the requirements of the contract," McCord explained.

Curtis McCord, Policy Analyst at the Canadian Anti-Monopoly Project

McCord emphasized that the government should see "competition as essential from the very outset" in its AI strategy. The CAMP report notes that hyperscalers maintain their dominance because they offer capabilities that competitors cannot yet match at scale. "The global reach, elastic scaling and rich ecosystem of platform services they provide represent decades of engineering investment and operational expertise that no other class of provider currently replicates at scale," the report states.

What Are the Broader Implications for Canada's AI Competitiveness?

Canada's challenge reflects a broader pattern: while the country has built a world-class AI research ecosystem since launching the Pan-Canadian Artificial Intelligence Strategy in 2017, the commercial gains from generative AI have flowed largely to US technology giants. Only about 12% of Canadian businesses used AI in 2025, a figure the government aims to significantly increase through its new strategy.

The government's approach attempts to address both supply and demand. On the supply side, it is investing in domestic compute infrastructure and supporting Canadian AI companies through equity stakes and procurement. On the demand side, it is targeting widespread AI adoption across the economy and providing free literacy training to build a skilled workforce. The strategy was originally scheduled for release before the end of 2025, then delayed to the first quarter of 2026, but Prime Minister Carney confirmed on May 27 that it would arrive the following week.

The core tension remains unresolved: building sovereign AI infrastructure requires competing against companies with decades of engineering investment and global scale. Canada's strategy bets that strategic government investment, combined with smart competition policy and interoperability requirements, can create space for domestic alternatives without simply replacing US hyperscaler dependency with Canadian telecom oligopoly dependency.