Hours after Defense Secretary Pete Hegseth abruptly ended Anthropic’s work with the Pentagon, President Donald Trump and his administration projected the dispute as a national-security test of whether leading AI technology should be routed into military systems. The clash surfaced publicly after Hegseth’s decision and the administration’s accusations that Anthropic’s products could be used in ways that threaten national security.
Hegseth’s move followed a standoff with Anthropic CEO Dario Amodei, according to the Associated Press report. Trump and Hegseth said Anthropic posed risks after Amodei declined to back away from concerns that the company’s products could be used for mass surveillance or autonomous armed drones, drawing a sharp response from the San Francisco-based company.
The administration also cast the action as grounded in law aimed at foreign supply-chain threats, and it treated the decision as an unprecedented step against a U.S. company. In the account, Anthropic said it was prepared to sue over Hegseth’s call to designate Anthropic as a supply chain risk, which the report described as “never before publicly applied to an American company.” Anthropic said it would challenge the decision in court, calling it legally unsound.
The Pentagon’s supply-chain risk designation, as described by the AP, is expected to end Anthropic’s contract with the Defense Department—reported as up to $200 million—and prevent other defense contractors from doing business with Anthropic. Trump also posted that most government agencies would stop using Anthropic, while the Pentagon would have six months to phase out technology already embedded in military platforms, according to the report.
Anthropic disputed key parts of that approach. The company argued that Hegseth did not have legal authority to stop business relationships with other defense contractors, and it said commercial contracts for non-defense projects could continue to use Anthropic’s products. Anthropic emphasized that the supply-chain designation, in its view, affects the use of Claude—the company’s AI chatbot and coding agent—for Department of Defense work, while other uses would remain unaffected.
Amodei, in a separate on-record interview that AP said would be aired on CBS, also made the company’s court challenge clear. “When we receive some kind of formal action, we will look at it, we will understand it and we will challenge it in court,” he said, according to the AP report. In the same reporting, Anthropic said it has not yet been formally notified of the designation.
The episode is also raising questions about how such restrictions could ripple through the broader tech and defense contracting ecosystem. The AP report said Anthropic is trying to reassure businesses and government agencies that the designation does not reach Claude’s use beyond defense contracting. That distinction matters, the report said, because a large share of Anthropic’s projected revenue comes from customers using Claude for computer coding and other tasks, including businesses and government agencies not tied to defense projects.
The report also described the competitive impact of the standoff inside AI markets, noting that hours after Anthropic’s punishment, OpenAI’s Sam Altman announced a deal with the Pentagon to supply OpenAI’s AI to classified military networks. Altman said that the same AI restrictions that were sticking points in Anthropic’s dispute with the Pentagon are now built into OpenAI’s partnership, including red lines around mass surveillance and autonomous lethal weapons. In a memo obtained by AP, Altman told OpenAI employees that the company believes “AI should not be used for mass surveillance or autonomous lethal weapons,” and that humans should remain in the loop for high-stakes automated decisions.
The AP report portrayed uncertainty around why the Pentagon accepted OpenAI’s stated red lines but pursued a supply-chain designation against Anthropic. It also framed the move as potentially deepening the rivalry between Altman and Amodei, noting their separate histories with OpenAI: Amodei left OpenAI in 2021 to launch Anthropic partly over AI safety concerns, and Altman was temporarily ousted by his board in late 2023 amid questions about his trustworthiness.
Beyond Anthropic and OpenAI, the dispute could also affect other companies seeking or preparing to work with the U.S. military on AI tools. The report cited Elon Musk’s public commentary in support of the administration’s posture toward Amodei and noted that Musk oversees the chatbot Grok, which the Pentagon plans to provide access to classified military networks. It also said Google’s leadership could face renewed pressure as the company has developed AI tools on its Gemini technology, while some employees have urged executives to avoid deals that would violate a former company motto.
The legal battle expected from Anthropic—and the Pentagon’s supply-chain action itself—could become a defining moment for how the U.S. government polices AI in military contexts. In AP’s account, Anthropic said it would challenge the designation in court and rejected the idea that it would change its stance on what the company described as the risks of mass domestic surveillance and fully autonomous weapons.
It was not immediately clear how quickly the six-month phase-out window will translate into operational changes for programs already embedding Anthropic technology, or how closely other contractors will weigh the risk of losing government business. For now, Anthropic’s focus is on contesting what it calls the Pentagon’s legal authority, while the administration is pushing a warning to companies that it says must meet its demands to work with the Defense Department.