Germany Bars Palantir From Military Cloud Over Data Rules

Germany has blocked Palantir Technologies from taking part in its military cloud project, citing concerns about data sovereignty and control by a foreign company. The armed forces have narrowed the field to three European firms — Almato and Orcrist of Germany and ChapsVision of France — whose software will undergo tests this summer. Officials expect to award a contract before the end of the year.

Vice Admiral Thomas Daum, inspector of Cyber and Information Space, said granting a private U.S. firm operational access to Germany’s national database is currently inconceivable and added he “does not see that happening right now.” The comment reflected a broader emphasis in the procurement process on keeping classified data and operational control within national or European hands.

The procurement covers a secure private cloud intended for data processing and artificial intelligence tools to support defense operations. German officials have described strict requirements for data access, operational oversight and hosting classified material as central to the selection criteria for the cloud software.

Palantir’s Maven platform already supplies intelligence outputs used by NATO and several member states, and Germany uses those outputs in some capacities. German officials drew a distinction between using intelligence products and allowing an external company to run or have operational control over a national database.

Political and ownership issues were part of the debate. Defense Minister Boris Pistorius previously raised questions about a minority stake held by investor Peter Thiel in Stark Defense, a German drone maker. That contract was approved only after assurances were provided that Thiel held no operational authority. Officials pointed to Palantir’s high political profile and ties to U.S. tech investors as complicating factors in evaluating the company’s role in the defense cloud.

The German decision comes amid an international focus on what policymakers call AI sovereignty. Some governments are seeking domestic or regional capacity for cloud and AI services to limit reliance on a small set of foreign providers. The United Kingdom has allocated funding for a sovereign AI unit, and countries including France and Brazil are developing regulatory frameworks aimed at keeping control of critical data and AI systems locally. Research from a U.S. university institute tracking global AI policy trends has documented a rise in such measures.

At the same time, a State Department cable instructed diplomats to lobby against foreign data sovereignty laws, arguing they could disrupt cloud and AI services and criticizing some regulatory approaches in Europe. European officials have in some cases responded by maintaining or strengthening local control measures.

The decision arrives as several U.S. AI-related companies prepare for public listings and continue to report high capital spending and operating losses. In 2025, one company’s AI division accounted for 61% of the firm’s $20.74 billion in total capital expenditure and reported an operating loss of $6.4 billion. Companies including that firm’s AI arm, OpenAI and Anthropic have said they do not expect to reach profitability before the end of the decade. OpenAI missed internal targets last year for weekly active users and annual revenue tied to its chatbot product; its chief financial officer warned the company could face difficulty funding future computing contracts without faster revenue growth, and some directors have questioned long-term data center commitments.

German officials say the cloud procurement will continue on the timeline for summer testing and a contract decision before year-end.