Pentagon says it has stopped using Anthropic's Claude AI, after missing its own late-August phase-out deadline
People familiar with the matter say Claude was still used in Iran operations as recently as last week; Google, xAI and OpenAI now hold Pentagon AI contracts.
A Pentagon official told the BBC the department "has ceased the use of Anthropic products." Defense Secretary Pete Hegseth labeled Anthropic a national security supply-chain risk in February 2026. He did so after the company refused to drop safety limits built into Claude, its AI model. Still open. Anthropic is suing over the designation.
Where Claude ran inside Pentagon systems before the ban:
- Claude ran inside the Maven Smart System, Palantir's intelligence and targeting platform, analyzing satellite, drone and signals data.
- Maven's five-year contract ceiling grew from $480 million to $1.275 billion, as planned users grew from hundreds to thousands.
- Anthropic was the first frontier AI company deployed on classified U.S. government networks, starting in June 2024.