A Concentrated Market, Documented in Detail
The Canadian Anti-Monopoly Project's June 2, 2026 report, Parting Clouds: Creating a Competitive Marketplace for Compute, puts a number on something Canadian IT buyers have grumbled about for years: Amazon (42%), Microsoft (31%) and Google (12%) together hold 85% of Canada's public cloud market — well above their 66% combined global share. Ottawa's own procurement illustrates the dependency. Since 2021, according to Canadian Press reporting picked up by Global News, the federal government has spent almost $1.3 billion on cloud services from U.S. companies, most of it going to Microsoft.
The report lands two days before Prime Minister Mark Carney's government answered, in part, with its own move: the June 4 launch of "AI for All," Canada's national AI strategy, which commits to building "sovereign compute and cloud infrastructure" alongside a public AI supercomputer. The timing isn't a coincidence — CAMP's report was explicitly framed as a warning ahead of that strategy, and the government's own text acknowledges parts of Canada's AI stack remain "under foreign control."
Steelmanning the Case for Intervention
CAMP's strongest argument isn't the market-share number itself — concentration in hyperscale cloud is a global phenomenon, not a uniquely Canadian one — it's the mechanism. Egress fees, proprietary APIs and bundled AI tooling create real switching costs that compound over a contract's life, and a government spending $1.3 billion with two vendors has limited leverage to negotiate on price or terms once workloads and data are embedded in a given platform. The report's warning about "maplewashed dependencies" — domestic cloud entrants that still run on foreign chips, foreign standards and foreign billing logic — is a fair rebuttal to sovereignty theater that changes the flag on the server rack without changing who controls the switching costs. And Canada isn't inventing this problem from scratch: the EU's Data Act, applicable since September 12, 2025, already mandates a maximum 30-day switching window, phases out switching charges entirely by January 2027, and requires providers to publish machine-readable interfaces for data portability. If a comparable market can absorb interoperability mandates without collapsing hyperscaler investment, that's real evidence the sky doesn't fall.
Where the Report Overreaches
But CAMP's seven-point plan bundles a narrow, defensible fix with a much larger institutional one, and the two shouldn't travel together. Banning egress fees and requiring portability standards in federal procurement is proportionate: Ottawa is the customer here, and a buyer can set its own contract terms without new legislation. Attaching interoperability conditions to the $1-billion-plus AI Compute Access Fund CAMP wants tied to domestic investment is likewise a procurement lever, not a regulatory one — cheap to implement, hard to abuse.
The call for a dedicated cloud regulator, however, invents a standing bureaucracy to police a market where Canada's existing competition law already has jurisdiction. The federal Competition Bureau has run digital-economy inquiries before — its 2022 call-out for information on anti-competitive conduct in digital markets covered search, social media and marketplaces — and CAMP's own report asks the Bureau to open a formal market study, which is the right first move. A sector-specific regulator, layered on top of the Bureau and CRTC, adds a second body with overlapping authority before anyone has established that ordinary competition enforcement — which can already probe bundling, self-preferencing and tying under the Competition Act — is inadequate to the task. Regulatory duplication is its own tax on the market CAMP wants to open up, borne disproportionately by smaller Canadian cloud entrants who can't afford compliance teams for two regulators instead of one.
The CUSMA Complication
There's also a live trade dimension the report gestures at but doesn't fully reckon with. Canada, the U.S. and Mexico are mid-way through the mandatory 2026 Joint Review of CUSMA, triggered on the agreement's sixth anniversary this July 1. CUSMA's digital trade chapter constrains exactly the kind of data-localization and platform-neutrality rules CAMP wants; unwinding those constraints unilaterally, mid-review, risks handing Washington a grievance to litigate rather than negotiate. Sequencing matters: portability mandates tied to Ottawa's own procurement contracts are safely within Canada's existing authority regardless of CUSMA's text. A binding domestic interoperability statute aimed at hyperscalers is the kind of measure that belongs in the CUSMA conversation, not ahead of it.
The Better Version of This Fix
CAMP has identified a genuine problem — a captive federal customer paying nine figures a year into a two-vendor market — with a genuinely proportionate first-order fix: use procurement leverage to demand portability, kill egress-fee lock-in on government contracts, and let the Competition Bureau's existing market-study power do the diagnostic work before Parliament creates a new regulator to solve a problem that hasn't yet been shown to require one.
Concentration is a symptom worth treating. A new standing bureaucracy is a much bigger commitment than the evidence, so far, justifies.