Sen. Warren Calls Pentagon's Blocking of Anthropic 'Retaliation' for Refusing Surveillance Tools

Senator Elizabeth Warren (D-MA) accused the Department of Defense of retaliating against Anthropic in a letter sent to Defense Secretary Pete Hegseth, according to a CNBC report. “I am particularly concerned that the DoD is trying to strong-arm American companies into providing the Department with the tools to spy on American citizens and deploy fully autonomous weapons without adequate safeguards,” Warren wrote, adding that the barring of Anthropic “appears to be retaliation.”

The timing is pointed. The letter arrived the day before a hearing in San Francisco, where District Judge Rita Lin is scheduled to decide whether to grant Anthropic an emergency preliminary injunction against the designation.

What the DOD designation actually means

The Pentagon’s “supply-chain risk” label is not a fine or a regulatory penalty. But it functions as one in practice. Any company or agency that does work with the Defense Department must certify it does not use products from designated entities. Applied to Anthropic — which counts government contractors, financial firms, and healthcare companies among its Claude API customers — the label effectively bars Anthropic from the commercial ecosystem that overlaps with federal work. That’s a large fraction of U.S. enterprise.

The dispute traces to Anthropic’s refusal to grant the Pentagon certain terms of use. Specifically, Anthropic told DOD it did not want Claude used for mass surveillance of Americans and that the model was not ready to make targeting or firing decisions for lethal autonomous weapons without human oversight. The Pentagon’s position: a private company should not dictate military use of technology it sells.

Growing support for Anthropic

Warren is not the first or only institution to push back on the Pentagon’s position. Tech companies including OpenAI, Google, and Microsoft have filed amicus briefs in support of Anthropic. Legal rights organizations have also joined, arguing that the designation mechanism — designed to protect against foreign adversary infiltration of supply chains — was never intended for application to U.S. companies refusing to enable domestic surveillance.

The breadth of that coalition is notable. OpenAI and Anthropic compete directly on foundation models. That OpenAI filed in support of its rival signals that the tech industry views the DOD’s action as a precedent threat, not just an Anthropic-specific problem. If the government can blacklist an AI company for setting ethical limits on military use cases, every AI lab faces the same exposure.

What the court will weigh

Judge Lin’s decision on the preliminary injunction will turn on whether Anthropic can show it faces irreparable harm and whether it has a likelihood of success on the merits. The supply-chain risk designation has already cost Anthropic deals and is chilling new ones. On the merits, the core question is whether the DOD exceeded its statutory authority by applying a foreign-adversary designation to a domestic company for reasons unrelated to supply-chain security.

The broader outcome matters beyond Anthropic. It will define whether AI companies can maintain ethical deployment standards when those standards conflict with government demands.