On April 15, 2026, the European Commission announced €1.07 billion for 57 “sovereign” defense AI projects under the European Defense Fund.
That same week, European military commanders were running live operations through Palantir’s systems.
That gap between the language in Brussels and the reality on the ground is Europe’s most consequential technology problem right now. This is not about consumer AI or market competition. This is about the infrastructure that determines whether European militaries can see, decide, and act in a war. And right now, that infrastructure is American.
NATO’s Command System Is a Palantir Product
In March 2025, NATO’s Communications and Information Agency finalized the acquisition of the Palantir Maven Smart System NATO (MSS NATO) for Allied Command Operations at SHAPE in Mons, Belgium. The procurement was one of the fastest in the Alliance’s history, taking just six months from requirement to contract. MSS NATO is already deployed at SHAPE and JFC Brunssum in the Netherlands, with JFC Norfolk currently in the process of receiving it.
MSS NATO is not a peripheral tool. It fuses satellite imagery, drone video, sensor data, and intelligence reports into a single interface that commanders use to see the battlefield and make decisions. General Markus Laubenthal, SHAPE Chief of Staff, described it saying: “Maven Smart System NATO enables the Alliance to leverage complex data, accelerate decision-making, and by doing so, adds a true operational value.“
The dependency extends well beyond NATO headquarters. The UK Ministry of Defense awarded Palantir a £240.6 million contract in December 2025, without competitive tender, under a national security exemption. As part of a broader strategic partnership, Palantir committed to invest up to £1.5 billion in the UK and establish London as its European defence headquarters.
France’s domestic intelligence agency, the DGSI, renewed its Palantir contract for another three years in December 2025, extending a relationship that dates to 2016.
Spain has also confirmed intelligence fusion contracts with Palantir, and Poland has signed agreements covering AI, data integration, and cybersecurity.
Palantir now sits within the command and intelligence architecture of multiple NATO member states.
Starlink and Ukraine Showed What “Dependency” Actually Means
Ukraine exposed what happens when a military relies on American-controlled infrastructure with no fallback. Starlink terminals became the essential backbone of battlefield communications, connecting command centers, enabling drone operations, and correcting artillery fire. Ukrainian officials have openly acknowledged their reliance on the system.
The critical detail is not that Starlink worked. It is that those operational decisions made inside SpaceX, a private company, shaped battlefield outcomes. In February 2025, U.S. negotiators reportedly threatened to restrict Ukraine’s access to Starlink during negotiations over a minerals deal. In February 2026, SpaceX worked with Ukraine’s defense ministry to enforce stricter authentication controls that blocked Russian military use of the network, creating an immediate crisis on Russian frontlines.
A private firm’s terms of service, not any treaty or parliament, governed what happened on the battlefield.
The same logic applies to Palantir. European militaries are now running command, intelligence fusion, and targeting processes through software hosted by an American company, subject to U.S. law, and ultimately accountable to U.S. interests. The EU AI Act’s military exemption means this system operates outside Europe’s own regulatory framework.
If Washington and Brussels diverge on Ukraine, tariffs, and technology regulation, that divergence is already visible, and the question of who controls the software layer of European military operations becomes operationally urgent.
Germany Said No. The Rest of Europe Hasn’t.
Germany has started pushing back. In April 2026, Bundeswehr cyber chief Vice Admiral Thomas Daum told Handelsblatt that Germany’s armed forces have no plans to award contracts to Palantir: “It is simply inconceivable at the moment to grant industry staff access to the national database.” Within weeks, Germany’s domestic intelligence agency, the BfV, awarded the French firm ChapsVision a contract for large-scale data analysis, an explicit signal against dependency on U.S. providers.
Three German states: Bavaria, Hesse, and North Rhine-Westphalia, still use Palantir’s Gotham software for law enforcement. And at the NATO level, Germany has not extracted itself from MSS NATO, which is embedded across Allied Command Operations.
The rest of Europe has not seriously asked the question Germany is asking. The UK deepened its Palantir relationship. France renewed one contract while separately praising sovereign AI companies like Mistral. The pattern is consistent: European governments acknowledge the dependency problem in official statements while extending the contracts that entrench it.
€1.07 Billion Will Not Close a €150 Billion Gap
The EU’s €1.07 billion EDF package is real money: 57 projects covering AI, cyber defense, drones, and counter-drone systems. The Commission has also proposed AGILE, a fast-track €115 million pilot instrument for disruptive defense startups. Europe’s defense tech sector is growing: Helsing AI, the German battlefield AI company, is currently in talks for a new funding round valuing it at approximately $18 billion, and the European defense tech deal count grew 67% in 2025 compared to 30% in the U.S.
But in May 2026, the Kiel Institute for the World Economy, which published Sparta 2.0, co-authored by former Airbus CEO Thomas Enders and four senior German defense figures, estimated that genuine European defense autonomy would require €150 billion to €200 billion by 2030 and €500 billion over the next decade. The paper identified command and control as the top strategic gap: Europe has no sovereign C2 software ecosystem, and no European combat mission is currently conceivable without U.S. software or systems. Closing the C2 gap alone would take three to four years and cost between €10 billion and €20 billion.
A €1.07 billion grant round funds research. It does not build the command-and-control architecture that European militaries would need to operate without American systems.
The Question Europe Keeps Avoiding
Brussels has framed its AI policy primarily around commercial competition and regulatory structure. That framing consistently misses what is happening in defense procurement. As Chatham House noted in April 2026, Palantir’s CEO, Alex Karp, has explicitly positioned his company’s mission as ensuring that the West continues to “dominate” technologically. The company’s April 2026 manifesto, based on Karp’s book ‘The Technological Republic’, states that Silicon Valley has “an affirmative obligation to participate in the defense of the nation.” That nation is the United States. Not European interests.
Consumer AI regulation is a policy debate. Defense AI dependency is a security question. Europe is treating them as the same problem, and until that changes, every “sovereign AI” announcement from Brussels is political theater performed on top of American infrastructure.
See Also:
Trump Took Silicon Valley to China. Where Was Europe?
