Trump Bans Anthropic from Federal Use; OpenAI Strikes Pentagon Deal Hours Later
On February 27, 2026, President Trump ordered all federal agencies to immediately cease using Anthropic's AI products. Defense Secretary Pete Hegseth simultaneously designated the company a "supply chain risk to national security"—effectively blacklisting it from military contracts in an unprecedented action.
The dispute centers on Anthropic's refusal to remove two safety guardrails from its proposed $200 million military contract. CEO Dario Amodei stated the company could not "in good conscience" agree to a deal without explicit prohibitions on two uses: domestic mass surveillance of American citizens and fully autonomous weapons systems. The company argued that current AI technology cannot safely handle these applications.
Hours after Trump's announcement, OpenAI CEO Sam Altman announced a competing agreement with the Pentagon. Altman said the deal includes the same two limitations Anthropic had demanded. However, the implementation differs: while Anthropic sought hard contractual language, OpenAI agreed that the DoD may use its AI for "any lawful purpose" while separately noting the restrictions in its own agreement language.
MIT Technology Review characterized the situation bluntly: "OpenAI's compromise is what Anthropic feared." The episode reveals a core tension in AI safety governance—whether safety constraints must be explicitly enshrined in binding government contracts, or whether AI companies' own internal policies provide sufficient protection.
Read more at MIT Technology Review and NPR.
Related Articles
The U.S. Department of Defense finalized AI deployment agreements with OpenAI, Google, Microsoft, AWS, NVIDIA, SpaceX, Reflection AI, and Oracle for its most classified networks. Anthropic was excluded after refusing to allow Claude to be used for purposes including autonomous weapons and mass surveillance.
OpenAI announced its EU Cyber Action Plan on May 11, granting vetted European security teams access to GPT-5.5-Cyber. UK AISI testing found GPT-5.5 scores 71.4% on expert cyber tasks, edging Mythos at 68.6%.
At the India AI Impact Summit held in New Delhi February 19–21, tech commitments surpassed $400B. OpenAI partnered with TCS for data center capacity, and Anthropic teamed with Infosys to deploy Claude across Indian enterprises.