Europe's De-Palantir Movement — The Litmus Test for Digital Sovereignty

When the heads of France's and Germany's security services announced last month that they plan to build a "European sovereign digital backbone," insiders in the technology and defense industries on both sides of the Atlantic knew exactly what it meant: "Goodbye, Palantir." The US-headquartered data analytics company — whose software has permeated everything from local policing to global intelligence, defense, and healthcare systems — now faces a continent-wide withdrawal campaign.

Crisis into Contract: Palantir's Foothold Model in Europe

Palantir's expansion path across Europe has been strikingly consistent — it has never missed a crisis. After the 2015 Paris terror attacks, security services under public-accountability pressure urgently needed tools that could show quick results, and Palantir promptly entered the French market; in 2016, France's domestic intelligence agency signed a contract with it. Its first major deployment in Germany followed the same pattern: the Frankfurt police purchased the "Gotham" system in 2017 and put it into operation under the name "hessenDATA," where it became a key tool for officers turning vast quantities of information into investigative leads. Its partnership with Britain's NHS began during the pandemic — at the time, Palantir charged a symbolic £1 to help consolidate data.

This "crisis into contract" model explains why Europe wants to quit but cannot: in moments of crisis, decision-makers want "something that works right now," not "something that is sovereignly correct but has not yet been built." Palantir's product capabilities themselves generate stickiness — Philippe Latombe, a French advocate of digital sovereignty, put it this way: Palantir's products are "excellent and addictive — it is almost like the sugar in Coca-Cola."

Attempts at Withdrawal, and the Real-World Friction

The wave of shedding Palantir is spreading across the continent: Spain's Sánchez government has instructed state-owned enterprises to exclude Palantir from future public procurement contracts; France's domestic intelligence service (DGSI) chose the French firm ChapsVision over Palantir; and Britain's test may come in February 2027, when the new Labour government will face the decision of whether to terminate Palantir's £330 million contract for the NHS's federated data platform.

But the friction is equally real. Admiral Pierre Vandier, NATO's Supreme Allied Commander Transformation, has stated plainly that the alliance currently has no viable alternative to Palantir's battlefield artificial-intelligence technology — which sifts through vast volumes of satellite imagery to help identify targets, suggest which weapons to use, estimate munitions requirements, and automatically place resupply orders. On July 1, Palantir's "Maven Smart System" went fully operational at NATO, granted security clearance to run on classified networks and tying together the command-and-control systems of the entire alliance.

Digital Sovereignty Presupposes Industrial Policy

Olivier Dellenbach, chief executive of ChapsVision, puts his finger on the point: he does not want ChapsVision to become a mere anti-Palantir convenience. In his view, unless governments translate "digital sovereignty" into industrial policy, it will remain an empty slogan — "We need more government procurement."

Belgium, Germany, Luxembourg, Romania, the Netherlands, and Canada have expressed interest in "Artemis," an AI solution developed by France's Thales, hoping to choose "sovereign solutions that meet NATO standards" — unlike Palantir's "black box." But Europe's market for alternatives remains fragmented, and European firms cannot yet match Palantir's scale or track record.

Even more paradoxical is the reverse move in capital markets: even as political leaders called for "decoupling" from Palantir, Europe's largest banks and asset managers significantly increased their holdings in the American company over the past year, as it positioned itself aggressively to profit from the AI boom. Withdrawal at the political level and doubling-down at the capital level happening simultaneously — that is perhaps the most honest footnote to Europe's digital-sovereignty predicament.

Incremental Update: The Other Side of Profit-Shifting — How Palantir "Compresses Its European Profits" (2026-08-06)

Beyond the withdrawal campaign runs a parallel economic thread: Palantir's tax structure in Europe. The Centre for International Corporate Tax Accountability and Research (CICTAR), a UK-based body partly funded by trade unions, published a report arguing that Palantir is shifting the profits of its European operations to the United States — its European subsidiaries recorded annual revenue of €440.5 million in 2024, yet reported profit margins far below those in the US.

Two sets of numbers lay out the gap clearly. In 2025, Palantir's US operations earned 47.7 cents of profit for every dollar of revenue — more than double the 22.5 cents of the previous year; outside the US, the margin was just 6.3%, and in some European subsidiaries it fell to around 3%. CICTAR considers this a "deliberate and artificial" compression of European profits: multinationals typically reduce their reported profits by charging subsidiaries for intellectual property, loans, or specialized expertise. Take Sweden: in 2024, Palantir reported revenue of €13.7 million and profit of only €1.1 million there, leaving a tax bill of just €424,000 at a 20% rate.

This structure is cut from the same cloth as the older cases of Apple and Microsoft: in 2024, the European Court of Justice ruled that Apple must pay Ireland €13 billion in back taxes; Microsoft's "Round Island One" Irish subsidiary avoided millions of euros in tax by claiming tax residency in Bermuda. Goudriaan, Secretary General of the European Federation of Public Service Unions, puts it bluntly — these companies "deprive public services of their sources of funding," and companies bidding for public contracts should disclose where their revenues, employees, profits, and taxes are located.

Palantir denies the tax-avoidance allegations, saying its profit allocation is "completely standard practice for large multinationals" and that its tax position in each jurisdiction reflects the level of economic activity there. But set this thread alongside the de-Palantir movement, and Europe is in fact facing a double bind: it can neither withdraw from Palantir's product capabilities, nor collect the taxes it should be paying.

Incremental Update: Britain's "Buy British" — Three Tests Facing the Burnham Government (2026-08-11)

The de-Palantir movement now faces its highest-profile test: the United Kingdom. On August 11, Lingshi Xiantan relayed a Politico report: the British government under Prime Minister Andy Burnham has vowed to "Buy British" AI to strengthen digital sovereignty, but a slate of imminent decisions will put that pledge to the test.

Three questions sit before the British government, with Palantir central to all of them. First, before the end of the year, the decision on whether to renew Palantir's £330 million contract to provide the UK's National Health Service (NHS) with a Federated Data Platform (FDP). Second, the issuance of the first Single Patient Record (SPR) contracts — a new digital record system that integrates patients' scattered health data; some British MPs have urged the government to exclude Palantir from the potential supplier list once the FDP project gets under way. Third, the Home Office is drawing up a plan to integrate police datasets, expected to cost £250 million a year, with Palantir currently bidding for the work.

Under intense political pressure from MPs, activists, and parts of the British tech community, the government has decided to reduce dependence on US foreign-tech giants and support domestic alternatives. But critics warn this could heighten transatlantic tensions and shut out advanced technology; supporters insist that procurement reform needs to land in practice, not just be talked about.