The White House directed federal agencies to cease using Anthropic’s AI products and gave the Pentagon a six‑month phaseout after weeks-long clashes over contract terms and military use limits. President Trump publicly accused Anthropic of jeopardizing troops; the company’s CEO Dario Amodei refused to remove contractual limits on mass domestic surveillance and fully autonomous weapons. The administration threatened to brand Anthropic a “supply-chain risk,” a designation that could block the company from defense-related commercial partnerships. Within hours of that escalation, OpenAI announced a classified-use agreement with the Pentagon. OpenAI CEO Sam Altman said the deal includes technical guardrails and commitments against certain uses while allowing the Department of War lawful access. The dueling outcomes expose universities and research centers to shifting vendor relationships for classified research, constrained collaboration choices, and a potential reordering of who supplies advanced models for defense-oriented labs and grants.