On February 21, 2025, President Donald Trump signed a presidential memorandum with a title that left no room for interpretation: Defending American Companies and Innovators from Overseas Extortion and Unfair Fines and Penalties. It directed the Office of the U.S. Trade Representative to reopen Section 301 investigations into foreign digital services taxes (DSTs) — the ring-fenced levies that France, the United Kingdom, Italy, Spain, Austria, and other governments impose almost exclusively on a small group of large American technology firms.
Three weeks earlier, on January 20, Trump had withdrawn the United States from the OECD/G20 Global Tax Deal, the two-pillar framework that was supposed to replace these unilateral measures with a coordinated international regime. Together, the two actions reset Washington's posture on cross-border digital taxation. The first major casualty came in late June 2025, when Canada rescinded its Digital Services Tax Act after a tense round of bilateral talks that had briefly halted broader trade negotiations.
What DSTs actually do
A digital services tax is a turnover tax — typically 2 to 5 percent of in-country revenues from online advertising, intermediation services, or user-data monetisation — that applies above high revenue thresholds calibrated to hit only the largest platforms. France's 3 percent DST, enacted in 2019, applies to firms with €750 million in global digital revenues and €25 million in French digital revenues. The UK's 2 percent DST uses similar thresholds. In practice, the affected universe consists almost entirely of Google, Meta, Amazon, Apple, Airbnb, Booking, and a handful of others.
DSTs depart from a century of international tax norms in three uncomfortable ways. They tax gross revenue rather than profit, so a firm operating at a loss still owes the levy. They apply to a hand-picked sector rather than the economy at large. And by construction, they fall on companies headquartered elsewhere — almost always the United States. The 2019 and 2020 USTR Section 301 reports concluded that the French, Austrian, Italian, Spanish, Turkish, Indian, and UK measures were "unreasonable or discriminatory," and the Biden administration agreed enough to negotiate a temporary truce conditional on the OECD Pillar One process delivering an alternative.
The OECD process that didn't
Pillar One was supposed to reallocate a share of the world's largest multinationals' residual profits to "market jurisdictions" — exactly the user-base countries that DSTs were trying to reach — in exchange for a standstill and rollback of the unilateral taxes. After more than five years of negotiation, the Multilateral Convention to implement Pillar One has not been signed by the United States, which holds an effective veto because the regime cannot function without American participation. By early 2025, capitals from Paris to New Delhi were openly preparing for the truce to collapse.
The Trump administration's January 20 executive order made the collapse official, declaring that the OECD deal had "no force or effect" in the United States and instructing Treasury to investigate "extraterritorial" provisions — including the Pillar Two undertaxed-profits rule — that could be used against American firms.
The pro-innovation case against DSTs
Our editorial position has long been that DSTs are poor policy on their own merits, independent of who imposes them. Three reasons stand out.
First, they distort. Taxing gross revenue in a low-margin business — cloud infrastructure, marketplace commissions, ad-tech intermediation — can exceed the underlying profit margin entirely, functioning as a punitive levy on the activity itself. Empirical work, including a widely cited 2019 Deloitte study commissioned by industry groups in France, found that the bulk of DST costs are passed through to small and medium-sized advertisers, app developers, and end consumers in the form of higher fees. Amazon, Google, and Apple all announced fee increases in DST jurisdictions explicitly tied to the new levies.
Second, they discriminate. A tax that by construction targets foreign-headquartered firms is exactly the kind of measure that international trade law, including GATT national-treatment obligations and Section 301 of the U.S. Trade Act of 1974, exists to address. The USTR's published findings — not seriously contested in substance by the European Commission — concluded that DSTs as currently designed are not facially neutral.
Third, they freeze. Sector-specific taxes lock in a snapshot of which business models count as "digital" at a particular moment, and they entrench the position of incumbents that can absorb the cost while raising the entry barrier for the next generation of platforms — including many now emerging from Europe, India, and Brazil.
Where Washington overreaches
None of this means the February 2025 memorandum is a model of restraint. Section 301 is a blunt instrument — last seen at full strength in the 2018–2019 tariff war with China — and its application to close allies raises real risks. Threatening retaliatory tariffs on French wine, Italian cheese, or Spanish olive oil over a digital tax dispute conflates unrelated sectors and invites a mirror response. The Canadian episode worked partly because the bilateral economic relationship is so asymmetric; the calculus is different with the EU, where collective retaliation through the Anti-Coercion Instrument is an explicit option.
A more durable settlement would do three things: restart serious work on a leaner version of Pillar One focused on the largest, clearly digital-economy firms; sunset existing DSTs in participating jurisdictions on a fixed timeline; and commit all sides — Washington included — to abandon extraterritorial enforcement against firms headquartered abroad.
American technology platforms are not above paying tax in the countries where they earn revenue. The question is whether that tax is imposed through coordinated, neutral, profit-based rules, or through discriminatory unilateral turnover levies enforced under the threat of tariff retaliation. The former is good policy. The latter is what the world ended up with — and what the next eighteen months will determine whether to fix.