February 16, 2026 Anthropic is reportedly clashing with the Pentagon over how its Claude AI models can be used in military operations. The dispute centres on U.S. defence officials pushing AI companies to allow their tools to be used for all lawful military purposes, while Anthropic is resisting broader permissions.
According to multiple reports, the U.S. military has pressed leading AI firms to loosen restrictions on how their systems are deployed, particularly in operational settings.
Anthropic appears to be the most resistant among major model makers. The company has positioned itself as a safety-focused developer and is reportedly seeking firm limits around uses such as fully autonomous weapons and large-scale domestic surveillance. That stance has drawn frustration from defence officials eager to accelerate AI integration across military functions.
The tension intensified following reports that Claude may have been used during a U.S. operation targeting Venezuelan leader Nicolás Maduro. While the precise role of the model has not been publicly confirmed, sources have suggested it may have supported intelligence analysis in real time. The Pentagon is now said to be reassessing its relationship with the company in the aftermath.
Anthropic has pushed back on claims it obstructed military operations, stating it has not discussed specific missions and that all deployments must align with its usage policies. The company maintains that its partnerships with government clients are governed by clearly defined safeguards designed to prevent misuse.
Other major AI companies are also navigating similar demands. Reports suggest the Pentagon has approached multiple firms, including OpenAI, Google and xAI, about expanding military access to their systems, particularly in classified environments. Some companies are said to be more flexible than others, though details remain limited.
For the Pentagon, the appeal is clear. AI models can rapidly process large volumes of data, offering potential advantages in fast-moving operational contexts. Defence leaders have repeatedly emphasized the need to accelerate AI deployment to maintain technological parity with geopolitical rivals.
Negotiations are ongoing between Anthropic and the Pentagon as well as between the Pentagon and OpenAI, Google and xAI.
