type: timeline_event
On March 23, 2026 — the eve of Anthropic's preliminary injunction hearing before Judge Rita Lin — Senator Elizabeth Warren sent a detailed letter to Defense Secretary Pete Hegseth calling the supply chain risk designation against Anthropic "retaliation" for the company's AI safety positions. Warren argued that if the Pentagon's actual concern had been Anthropic's refusal to meet contract requirements, the department had a straightforward remedy available: it could have simply terminated the contract. Instead, by deploying a supply chain risk designation, the Pentagon had chosen the most destructive option available, one that barred Anthropic from all federal procurement and signaled to the entire technology industry that maintaining AI safety standards would be punished.
Warren's letter drew heavily on the CDT revelations from earlier in the month, expressing concern that the DOD was "trying to strong-arm American companies" into participating in bulk surveillance of American citizens and developing fully autonomous weapons systems without meaningful human oversight. She noted that the Pentagon had never publicly explained why Anthropic's specific safety objections — particularly around bulk commercial data collection and autonomous targeting — were unreasonable, instead characterizing the company's positions in vague terms like "bad faith" and "arrogance."
The timing of the letter was strategic, arriving the day before what promised to be the most consequential hearing in the case. Warren's intervention added congressional weight to Anthropic's legal arguments and put additional political pressure on the administration to justify its actions in a forum where it would face cross-examination. CNBC reported that Warren was coordinating with several other Democratic senators who were preparing their own oversight actions related to the Pentagon's AI procurement practices, suggesting that the Anthropic case was becoming a broader vehicle for congressional scrutiny of the administration's approach to military AI governance.