A US appeals court ruled on Friday that the Pentagon may keep Anthropic on a national security blacklist. This decision reverses a previous ruling from August, in which a federal judge had declared the designation illegal and baseless. The company is now officially classified as a national security supply-chain risk, which restricts its ability to participate in military contracts.
The conflict began in late February following a meeting between Defense Secretary Pete Hegseth and Anthropic CEO Dario Amodei. The Pentagon demanded that Anthropic remove restrictions on the use of its Claude models for lethal autonomous warfare and mass surveillance. Anthropic refused, citing insufficient safety testing for such applications, which jeopardized a $200 million relationship.
Following the dispute, President Donald Trump ordered federal agencies to cease using Anthropic technology. Consequently, the State Department migrated its StateChat system to OpenAI, while the Treasury and the Department of Health and Human Services began phasing out Claude. This marks the first public use of this procurement statute against a US firm.
The legal battle highlights a growing tension between AI developers and the military regarding control over technology. While OpenAI has secured a contract to deploy AI on classified networks with specific usage bars, it has publicly opposed the blacklisting of Anthropic. The ruling leaves billions of dollars in potential government business for Anthropic in question.
