The Rules Washington Is Writing
Whose Law Controls Canadian Data?
The Rules Washington Is Writing: Whose Law Controls Canadian Data
By Prof. Barry Appleton | Appleton’s CLAUSE & EFFECT Substack| 9 June 2026 · Part 1 of 2
This is the first of a two-part series on Canada’s digital sovereignty under the USMCA-CUSMA review. It follows a three-part series on the new Section 301 architecture: “Reciprocal in Name Only” (June 6), “Already in Motion” (June 7), and “Settled Out of Congress” (June 8). The analytical foundations for this discussion are in Code Before Clause, Locked In and Locked Out, and Algorithmic Empire and the New Digital Colonialism. Part Two, “The Code Canada Has Not Written,” publishes tomorrow.
The United States is not just negotiating tariffs. It is writing the rules of the global digital economy through trade instruments. While Canada talks about the upcoming Review daily, Canada’s new AI strategy walked directly into the middle of that project, without a legal shield.
The current U.S. administration took office on January 20, 2025. Within days, it had signed two executive orders on artificial intelligence that together define a governing philosophy: one eliminating the safety requirements Biden had placed on large AI models, and the other identifying AI data center infrastructure as a strategic imperative requiring vastly greater American energy generation.1 Neither framed these as technology policy choices. Both framed them as national security decisions.
That framing matters because it is the same framing now reaching Canada’s trade file. The November 2025 National Security Strategy declares that “maintaining American economic and technological preeminence is the surest way to deter and prevent a large-scale military conflict” and frames the Western Hemisphere as a zone in which the United States must be “preeminent,” including through identifying and protecting strategic resources and discouraging allied collaboration with non-Hemispheric competitors.2 AI is not peripheral to this strategy. It is the strategy.
The United States is not negotiating tariffs. It is writing the rules of the digital economy, and calling the result trade policy.
What the strategy actually requires
The administration’s AI infrastructure executive order is frank about the physical challenge: AI is the first digital service in modern life that requires America to build vastly greater energy generation capacity, and American energy capacity has stagnated since the 1970s, while China has rapidly built out its grid.2 That observation has a direct implication for Canada. Canada’s cold climate reduces data centre cooling costs by 30 to 40 percent. Canada’s electricity grid is approximately 82 percent emissions-free. Canada’s landmass borders the United States and is connected to it by infrastructure that took generations to build.3 The administration’s own documents describe Canada as a strategic resource for the AI infrastructure build-out, it says is essential to national security.
And yet the administration’s trade policy treats every Canadian measure designed to give Canada a governance role in that infrastructure as a trade barrier. The 2026 U.S. National Trade Estimate Report flags Canada’s sovereign cloud initiative as a procurement concern. The new U.S. Agreements on Reciprocal Trade (ARTs) with other states contain equivalent-restrictive-effect clauses that require signatories to mirror U.S. technology restrictions on third countries. The nine ART texts already prohibit digital services taxes across the board.
U.S. Trade Representative Greer has announced that alleged discrimination against U.S. technology companies and digital goods will be a Section 301 enforcement target.4 The Canada chapter of the NTE 2026 lists Canada’s Online Streaming Act, Online News Act, and Quebec’s algorithmic discoverability requirements as measures under monitoring.5
The contradiction is structural, not accidental. The United States wants Canada as a supplier of trusted AI infrastructure while simultaneously using trade instruments to prevent Canada from governing the digital economy in which that infrastructure operates.
You cannot use an ally’s land for your data centres while prohibiting that ally from writing rules for the data.
How the rules are being written in practice
As I explained in an earlier Substack, “Reciprocal in Name Only,” the Agreements on Reciprocal Trade are not mainly tariff agreements. They are governance agreements that lock in the future policy making capacity of treaty parties. The seven recurring mechanisms documented in the nine ART texts, forced-labour exclusions, equivalent-restrictive-effect clauses, restrictions on third-country-controlled firms, export-control alignment, investment screening, tightened rules of origin, and penalty clauses, are not about lowering tariffs between the parties. You can call it what you like, but they are about aligning the smaller party’s domestic regulatory architecture with American policy choices.
The equivalent-restrictive-effect clause is the clearest example. It does not ask the signing country to negotiate a tariff. It asks the signing country to adopt domestic law that mirrors American restrictions on third countries whenever the United States imposes such restrictions. That is not trade policy. That is foreign policy implemented through a bilateral trade instrument, signed by an executive without congressional approval, and enforced by a tariff threat that can be reimposed at any time.
The NTE 2026 shows how this works in practice for allied countries with their own digital ambitions. Japan’s 2024 subsidy program for domestic GPU purchases and sovereign AI infrastructure, designed to allow Japan to participate in the AI economy without total dependence on U.S. hyperscale cloud providers, was specifically identified in the NTE 2026 as inhibiting U.S. hyperscale cloud providers and creating an “uneven playing field.”6 Japan was doing exactly what Canada’s AI for All strategy proposes. The U.S. response, flagging it as a trade barrier, tells Canada precisely what to expect.
Why Canada’s AI for All walked into this
Prime Minister Carney launched Canada’s national AI strategy on June 4, 2026, two days before the USMCA-CUSMA Joint Review window opened.7 Its sovereignty pillar commits to building a world-leading public AI supercomputer and investing in “sovereign compute and cloud infrastructure, under Canadian governance,” with government procurement as the funding mechanism. Its trust pillar commits to an online safety regime for chatbots and social media users, privacy protections against surveillance pricing, and the Canadian AI Safety Institute evaluating AI models transparently.
These commitments are correct. They describe how a serious, sovereign digital power manages AI within its own territory. But every one of them maps directly to a category USTR has identified as a potential trade barrier, a Section 301 target, or a measure under monitoring. The sovereign cloud commitment formalizes what the NTE already treats as a procurement concern. The online safety regime is in USTR’s announced investigation category of “discrimination against U.S. technology companies and digital goods.” The AI model evaluation commitment is the kind of algorithmic governance measure I analyzed in Locked In and Locked Out as sitting in tension with USMCA-CUSMA’s data-flow and IP provisions.
I have been writing about this collision since August 2025. “The Digital Hinge of Sovereignty” (Clause and Effect, July 23, 2025) argued that the US-EU Framework Agreement was the opening of a new era in which digital governance and trade policy would collide. The five National Post columns from September and October 2025, on algorithmic empires, digital sharecroppers, the railway of the future, Canada’s surrender to foreign code, and the oversight of platforms that control what Canadians see, documented the structural dependency that makes Canada vulnerable to exactly this pressure. Those columns described the problem. AI for All describes the right ambitions. What is missing is the legal sequence.
Canada announced the right ambitions in the wrong order. The strategy is correct. The sequence is not.
Why the ally argument cuts the other way
When I testified before the U.S. Trade Representative in December 2025, I made an argument that the hearing record confirms but that Ottawa has not yet deployed publicly: a robust and independent Canadian digital governance framework is a U.S. national security asset, not a trade barrier.
The 2025 National Security Strategy instructs officials to work with hemispheric partners to identify and protect strategic resources and to roll back non-hemispheric infrastructure that embeds espionage and cyber risk.8 A Canada that retains the legal and institutional capacity to exclude high-risk telecom and cloud providers, audit AI systems, and enforce its own cybersecurity standards provides exactly the trusted hemispheric capacity the NSS calls for. A Canada whose regulatory autonomy has been negotiated away through ART commitments cannot make those exclusions credibly, because it will have agreed to equivalent-restrictive-effect obligations that bind its domestic policy to U.S. decisions rather than its own risk assessments.
The United States should therefore be pleased to have strong and independent Canadian digital systems, because those systems supply America.
In my USTR testimony, I put it this way: technological non-alignment, the right of each nation to participate in global AI ecosystems without subordination to any single digital power, should be embraced as compatible with American leadership. Pluralistic, interoperable governance makes North America stronger and safer. It avoids technological monoculture and spreads innovation capacity across all three partners.8
That argument has not yet been made systematically in the USMCA-CUSMA context. The Joint Review is the place to make it.
A Canada that cannot govern its own AI systems cannot be the trusted ally the National Security Strategy says it needs.
The question underneath the strategy
AI for All is necessary. It is also, as published on June 4 with no implementing legislation in sight and no Article 32.2 essential security invocation accompanying its commitments, a target list for the administration’s lawyers.
The question underneath the strategy is not whether Canada should pursue digital sovereignty. It is whether Canada can pursue digital sovereignty within its existing treaty framework, and if so, how.
This important question is the subject of the companion piece, “The Code Canada Has Not Written,” which examines what the legal sequence actually requires, where USMCA-CUSMA permits it and where it constrains it, and what Ottawa would need to do before July 1 to give AI for All any protection at the Joint Review table.
· · ·
Prof. Barry Appleton is Managing Partner of Appleton & Associates International Lawyers LP, Co-Director of the Center for International Law at New York Law School, and Interim Director of the Balsillie Legal Advisory Centre at the Balsillie School of International Affairs.
© 2026 Barry Appleton. All rights reserved.
The White House, Executive Order on Removing Barriers to American Leadership in Artificial Intelligence (Jan. 23, 2025) (revoking President Biden’s Executive Order 14110 and eliminating federal AI safety requirements); The White House, Executive Order on Advancing United States Leadership in Artificial Intelligence Infrastructure (Jan. 2025) [hereinafter “AI Action Plan”] (identifying the American electric grid as facing “historic constraints” from data centre and AI computing demand; declaring that “AI is the first digital service in modern life that challenges America to build vastly greater energy generation than we have today”; and noting that “American energy capacity has stagnated since the 1970s while China has rapidly built out their grid”).
The White House, National Security Strategy of the United States of America (Nov. 2025) [hereinafter “2025 NSS”] (declaring that “maintaining American economic and technological preeminence is the surest way to deter and prevent a large-scale military conflict”; calling for “preserving and growing America’s financial sector dominance”; and framing the Western Hemisphere as a zone in which the United States must be “preeminent,” including through identification and protection of strategic resources and discouragement of allied collaboration with non-Hemispheric competitors).
Barry Appleton, Know Your Ground: Canada’s Strategic Red Lines for the 2026 CUSMA Joint Review, SSRN Working Paper No. 6643319 (revised Apr. 30, 2026) (noting that Canada’s cold climate is estimated to reduce data centre cooling costs by 30% to 40%, and that the Canadian grid is approximately 82% emissions-free, making Canada strategically valuable for U.S. AI infrastructure requirements as identified in the AI Action Plan).
Reuters, USTR Expects New Section 301 Probes to Cover Most Major Trading Partners (Feb. 20, 2026) (reporting Ambassador Greer’s statement that new Section 301 investigations would cover discrimination against U.S. technology companies and digital goods and services); Dylan Moroses, USTR Seeking ‘Outcomes’ on DSTs, Stronger USMCA Rules, Law360 (Apr. 22, 2026) (reporting that nine reciprocal trade agreements already contain digital services tax prohibitions and that further 301 actions are drafted and ready).
Office of the U.S. Trade Representative, 2026 National Trade Estimate Report on Foreign Trade Barriers (Mar. 31, 2026), Canada chapter (identifying Canada’s Online Streaming Act, Online News Act, Shared Services Canada sovereign cloud RFI, and Quebec’s Bill 109 as measures under monitoring for USMCA-CUSMA implications); Barry Appleton, “USTR’s Opening Brief: How to Read the 2026 National Trade Estimate Report Before It Reads Canada” (Clause and Effect, April 8, 2026).
Office of the U.S. Trade Representative, 2026 National Trade Estimate Report on Foreign Trade Barriers (Mar. 31, 2026), Japan chapter (identifying Japan’s 2024 subsidy program for domestic GPU purchases and sovereign AI infrastructure as “inhibiting U.S. hyperscale cloud providers from participating in Japan’s strategic AI development initiatives and creating an uneven playing field”).
Prime Minister of Canada, Prime Minister Carney Launches AI for All: Canada’s New National Artificial Intelligence Strategy (June 4, 2026).
Barry Appleton, December 3, 2025 Testimony, Provisional Transcript at 207-209, Docket No. USTR-2025-0004 (arguing that services and digital trade are “the most dynamic part” of the U.S.-Canada economic relationship, and that technological non-alignment and robust Canadian data sovereignty “are not a threat to the United States but a prerequisite for a resilient partnership”); see also Barry Appleton, “Law, Not Leverage: A Rules-Based Path for the USMCA Review”, Rebuttal Submission, Docket No. USTR-2025-0004 (Dec. 12, 2025) (arguing that “technological non-alignment, the right of each nation to participate in global AI ecosystems without subordination to any single digital power, should be embraced as compatible with American leadership”)