Speed was of the essence when OpenAI CEO Sam Altman announced a Pentagon agreement late Friday, just hours after President Trump issued an order banning all federal agencies from using Anthropic’s AI technology. The rapid sequence of events suggests the deal may have been in progress before the public confrontation with Anthropic reached its dramatic conclusion.
Anthropic had been in tense negotiations with the Pentagon over the terms under which its Claude AI system could be used by the military. The company insisted on two firm limits — no use for autonomous weapons systems, no use for mass domestic surveillance. These were presented as ethical bright lines that the company would not cross regardless of commercial consequences.
The Trump administration framed these limits as unacceptable and unconstitutional, with the president personally condemning Anthropic on Truth Social in sweeping terms. The federal ban that followed effectively ended any near-term possibility of government contracts for the company, which has positioned itself as the most safety-focused major AI developer.
Altman’s Pentagon deal came with assurances that OpenAI shares Anthropic’s views on the key ethical points, and he called on the government to standardize these principles across all AI contracts. He also sent an internal memo acknowledging that the Anthropic situation had become an industry-wide issue, not just a bilateral dispute.
Nearly 500 employees from OpenAI and Google had already signed a joint letter backing Anthropic before Altman’s announcement, complicating the optics of the deal. The letter warned that the Pentagon was attempting to leverage competition between AI companies as a tool to extract compliance with demands the industry had collectively resisted.