The Hegseth-Anthropic dispute was never primarily about which AI model is technically best. It was about who controls the conditions of AI use inside the Defense Department — and Anthropic's refusal to yield that control is what produced the ban. The Pentagon awarded Anthropic a $200 million contract, then demanded unrestricted access that Anthropic's own safety commitments made impossible to grant. When negotiations collapsed, the administration moved to enforce compliance through contract exclusion. What it did not build was a transition path.