Pentagon Formally Labels Anthropic a "Supply-Chain Risk," Company Vows Legal Challenge

After weeks of public friction and high-stakes ultimatums, the US Department of Defense (DoD) officially labeled Anthropic a “supply-chain risk.” This designation, typically reserved for foreign companies linked to adversarial nations, marks the first time an American tech firm has received such a label. In response, Anthropic CEO Dario Amodei has made it clear: the company is heading to court.

Anthropic AI labeled a “Supply-Chain Risk” by US Pentagon

The conflict centers on a fundamental disagreement over “red lines.” Anthropic, which markets itself as a safety-first company, refused to let the military use its Claude AI models for two specific purposes: mass surveillance and fully autonomous lethal weapons. On the other hand, the Pentagon demanded unrestricted access for “all lawful purposes.” However, Anthropic held firm, arguing that a private company should maintain some ethical oversight over how its technology is deployed.

The label is far more than a slap on the wrist. This designation technically bars defense contractors from using Anthropic’s products in their government work. Defense Secretary Pete Hegseth previously suggested an even broader application, hinting that any company doing commercial business with Anthropic could face contract cancellations.

However, the timing of this decision is particularly striking. While the administration has set a six-month deadline to remove Claude from federal systems, the AI is reportedly currently integrated into active military operations. Following recent U.S. strikes in Iran, sources indicated that Claude-powered intelligence tools played a major role in the mission’s success.

Anthropic to challenge in court

Dario Amodei quickly addressed the situation, calling the designation “legally unsound.” In a formal statement, he clarified that Anthropic plans to challenge the DoD’s decision in federal court. Amodei’s argument hinges on the idea that the law requires the government to use the “least restrictive means necessary” to protect the supply chain. He contends that the Pentagon is using the law to punish a supplier rather than to actually protect national interests.

Amodei also sought to reassure commercial clients. He stated that the vast majority of Anthropic’s customers remain unaffected. After all, the designation cannot legally limit business relationships that are unrelated to specific Department of War contracts.

The legal path won’t be easy. Experts, including former White House advisers, note that courts are traditionally reluctant to second-guess the government on matters of national security. The Pentagon enjoys broad discretion in these cases, meaning Anthropic faces a very high bar to overturn the label.

Meanwhile, the void left by Anthropic is already being filled. Rival OpenAI has reportedly signed a deal with the Pentagon to deploy its models in Anthropic’s place—a move that has sparked internal backlash among some OpenAI staff.