Overview
- Defense Secretary Pete Hegseth is described as close to ending Pentagon business with Anthropic and considering a supply‑chain risk designation, according to a senior official quoted by Axios.
- Such a label would force defense contractors to certify they do not use Claude, with a Pentagon contract worth up to $200 million also at risk of cancellation.
- Claude is currently the only AI model available on classified U.S. military systems and was reportedly used in the January operation targeting Nicolás Maduro’s regime, making any switch complex.
- Pentagon negotiators seek an “all lawful purposes” standard for AI use, while OpenAI, Google and xAI have already lifted safeguards for unclassified systems and are expected by officials to accept broader terms.
- Anthropic says it supports U.S. national security yet will not enable mass domestic surveillance or fully autonomous weapons, arguing existing laws have not kept pace with AI’s capability to analyze public information at scale.