Anthropic vs. Pentagon: Court Filings Reveal DoD Privately Said Sides Were "Nearly Aligned"
A surprising piece of the Anthropic-Pentagon dispute came into sharper focus this week as sworn court declarations painted a very different picture than the one the DoD presented publicly. Anthropic's Head of Policy Sarah Heck and Head of Public Sector Thiyagu Ramasamy both filed statements attesting that a Pentagon official told the company the two sides were "nearly aligned" on a revised agreement — and that was just one week before Trump and Defense Secretary Pete Hegseth announced they were cutting ties entirely over Anthropic's refusal to allow unrestricted military use of Claude.
Heck's declaration is particularly pointed. She flatly refutes the DoD's central allegation — that Anthropic had demanded an approval role over specific military operations — and states that the claim never came up during what had been months of substantive negotiations. The filings suggest the public break was less a negotiation failure and more a deliberate decision, made at the political level, to label Anthropic a "supply-chain risk" rather than continue working through the outstanding issues on the table.
The stakes here go well beyond one company's government contract. How this case is resolved will effectively set the rules for whether AI developers can enforce safety-use policies on defense customers — or whether federal agencies can override those limits by invoking national-security language. A hearing before U.S. District Judge Rita Lin is scheduled for March 24th, and the outcome could reshape how every major AI lab approaches defense contracting going forward.