US software company Palantir has said it cannot remotely shut down a military artificial intelligence system used by NATO, responding to European concerns about reliance on American defence technology. The statement follows broader debate about whether critical military systems could become vulnerable to political disputes between allies.
The system, known as Maven Smart System, processes battlefield information from different sensors to help military personnel assess situations and make decisions. NATO has used it at its military headquarters since mid-2025. The software is designed to bring together information that may otherwise be spread across separate systems, helping users analyse developments more quickly.
Palantir’s executive vice-president for the UK and Europe, Louis Mosley, said customers retain control over the software and that neither the company nor the US government has an external switch that can simply turn it off. The statement was intended to reassure European officials concerned about the risks of dependence on technology supplied by a foreign company.
The concerns are part of a wider discussion about digital sovereignty. Governments increasingly rely on privately developed software for communications, logistics, intelligence and defence. These systems can offer advanced capabilities, but they also raise questions about data access, maintenance and the ability to continue operating if political relationships deteriorate.
Recent problems involving other satellite services used by Ukraine have added to the debate. When military operations depend on digital systems, any disruption can affect communications and planning. European governments therefore want to understand what control they retain over systems provided by external companies.
Not all NATO members have adopted Maven. Germany and France have reportedly held back because of concerns about data security and political dependence. Their decisions reflect different views about how quickly advanced software should be introduced into military operations and what safeguards are required.
AI systems can help military users process large amounts of information, but they also require human oversight, secure data handling and clear rules about responsibility. A system’s technical ability to remain operational does not by itself settle questions about how its outputs should be used or who is accountable for decisions.
The issue is likely to remain important as NATO modernises its technology. Governments must balance the benefits of advanced tools against the need for reliable control, security and independence. Palantir’s assurances address one concern, but European officials will continue to examine the wider risks of relying on foreign technology for sensitive defence work.
