Overview
- The Defense Department said the designation is effective immediately, requiring defense contractors to certify they are not using Anthropic’s models in Pentagon work as agencies move to cut or shift usage.
- Anthropic said it will challenge the action in court and emphasized the label applies only to DoD-related contracts, with Microsoft, Google and AWS indicating Claude remains available for non-Defense customers.
- Reports say CEO Dario Amodei resumed talks with Pentagon officials to seek a resolution or managed transition, though the department’s chief technologist posted that no active negotiation is underway.
- News outlets report Claude has supported intelligence analysis and planning tied to U.S. operations against Iran, and its integration via partners such as Palantir complicates a rapid disentanglement.
- OpenAI moved to replace Anthropic in classified settings and later revised its agreement after backlash, while contractors like Lockheed Martin said they will comply as lawmakers and former officials criticized the designation as an unusual and risky precedent.