Anthropic and the Pentagon are said to be at loggerheads over the use of Claude

The Pentagon is pushing AI companies to allow the U.S. military to use their technology “with full approval,” but Anthropic is pushing back, according to A new report in Axios.

The government says it wants the same from OpenAI, Google, and xAI. An anonymous Trump official told Axios that one of the companies has agreed, while the other two are showing flexibility.

Anthropic, meanwhile, is said to be the most critical. In response, the Pentagon appears to be threatening to pull the plug on its $200 million contract with the AI ​​company.

In January, The Wall Street Journal said that there was considerable disagreement between Anthropic and Defense Department officials about how Claude’s models could be used. The WSJ later said Claude served in the US Army to arrest the President of Venezuela at the time Nicolás Maduro.

Anthropic did not immediately respond to TechCrunch’s request for comment.

A spokesperson for the company told Axios that the company “has not discussed the use of Claude for specific purposes with the Department of Defense” but instead “focuses on specific questions of the Use Policy — namely, our limits against autonomous weapons and mass surveillance at home.”