Pentagon pressures Anthropic to drop limits on Claude for military use
Feb 28th 2026
After Anthropic limited Claude’s use to block mass domestic surveillance and autonomous lethal decisions, the Defense Department threatened forced access under the Defense Production Act or a supply chain ban, raising legal, industry and national security risks.
- Anthropic set two contract limits: no mass domestic surveillance and no autonomous kill decisions without human authorization.
- Anthropic put Claude on classified networks and supplied customized models to national security customers.
- Pentagon officials have threatened to use the Defense Production Act or declare Anthropic a supply chain risk if the company will not permit any lawful military use of Claude.
- Leaked briefings and an ultimatum reported after a meeting between CEO Dario Amodei and Defense Secretary Pete Hegseth escalated the dispute into public conflict.
- Anthropic says its models are not ready for lethal autonomous weapons and many employees oppose building such capabilities now.
- Experts warn the fight could deter other AI firms from working with the U.S. military, prompt litigation, and harm Anthropic’s business and investor plans.