What changed
The Pentagon says it has stopped using Anthropic’s AI products, completing a break it had announced for late August. Claude had reportedly still been used as recently as the previous week, including through Palantir’s Maven Smart System, the Pentagon’s main platform for organising intelligence and other data.
Why it matters
This is less like deleting an app than removing a component from a large machine. Anthropic’s tools had been used by the US government and military since 2024 and were integrated into Maven, so ending their use required more than switching a contract.
The split began after Defence Secretary Pete Hegseth designated Anthropic a national-security supply-chain risk earlier this year. The Pentagon had pressed the company to remove safety restrictions and give the military unfettered access; Anthropic refused, citing concerns about mass surveillance and autonomous weapons.
The Pentagon already has contracts with Google, xAI and OpenAI, and OpenAI tools have become more widely used in some military departments. The immediate practical question is whether those alternatives can replace Claude inside existing systems without disrupting the Pentagon’s research, analysis and intelligence work. The timing and extent of any remaining transition are unclear.
Comments
No comments yet.