The growing clash between the U.S. Department of Defense and artificial intelligence company Anthropic has entered a decisive moment. At the center of the standoff is Pete Hegseth, who has delivered a firm deadline to the San Francisco–based AI startup. By 5 p.m. Eastern Time on Friday, Anthropic must agree to revised Pentagon terms governing how its technology can be used by the military — or risk losing its federal contract and potentially facing sweeping government action.
The dispute underscores a broader tension shaping the future of artificial intelligence in national security: how far private companies can — or should — go in limiting how their technologies are deployed once they enter government hands.
A High-Stakes Deadline
The ultimatum was delivered during a recent meeting between Hegseth and Anthropic’s chief executive, Dario Amodei. According to individuals familiar with the discussion, the Defense Department wants Anthropic to loosen restrictions tied to its flagship AI model, Claude, allowing broader military application without company-imposed constraints.
The Pentagon has reportedly made clear that if Anthropic declines to comply, several consequences could follow. Among them: termination of an existing contract valued at up to $200 million, a formal designation of the company as a supply chain risk, and possible invocation of the Defense Production Act.
The Defense Production Act, first enacted during the Cold War, grants the federal government authority to direct private companies to prioritize national defense needs during emergencies. It has historically been used to increase production of critical materials, including medical supplies during the COVID-19 pandemic. Applying it to an AI software company in order to compel cooperation would represent an extraordinary step.
A senior Pentagon official indicated that if Anthropic does not agree to the department’s terms by the specified deadline, the Defense Production Act could be triggered almost immediately afterward. The official also suggested that the supply chain risk label would be applied simultaneously.
The Core Disagreement: Guardrails vs. Military Flexibility
At the heart of the confrontation lies a philosophical divide over the responsible use of artificial intelligence in defense operations.
Anthropic has built its reputation around safety-oriented AI development. The company has publicly maintained strict usage policies that prohibit certain applications of its systems. Among its red lines are mass domestic surveillance of U.S. citizens and the use of AI in fully autonomous weapons systems where machines make lethal targeting decisions without direct human control.
Amodei is said to have reaffirmed those principles in his conversation with Hegseth, making clear that Anthropic is unwilling to remove those boundaries.
The Pentagon, however, argues that its use of advanced technologies is already governed by U.S. law and established military oversight procedures. Defense officials maintain that AI systems used by the armed forces must be available for all lawful purposes tied to national security. From their perspective, allowing a private company to restrict certain forms of military use could hinder operational readiness and create vulnerabilities.
The tension reflects a growing reality: as AI becomes more central to intelligence analysis, logistics planning, battlefield decision support, and cybersecurity, disagreements over ethical constraints are no longer theoretical. They now carry direct operational consequences.
Spotlight on Claude’s Military Use
Public scrutiny intensified earlier this year after reports that Claude had been used during a January U.S. military operation that resulted in the capture of former Venezuelan President Nicolás Maduro.
While details of the operation remain limited, the disclosure highlighted just how embedded commercial AI systems have become within defense workflows. It also illustrated why the Pentagon may be reluctant to risk losing access to Anthropic’s technology at a time when global competition over AI capabilities is intensifying.
For critics, however, the episode raised questions about transparency and oversight. If AI tools are playing meaningful roles in sensitive military missions, they argue, the public deserves clarity on how those systems are governed and what safeguards are in place.
Anthropic’s Public Position
Anthropic confirmed that discussions between Amodei and Hegseth took place. In a statement, company spokesperson Maya Humes described the talks as ongoing and conducted in good faith, emphasizing Anthropic’s desire to support national security while remaining aligned with what its models can “reliably and responsibly” do.
Notably absent from the company’s response was any indication that it intends to abandon its ethical guidelines.
Since its founding in 2021 by former OpenAI researchers, Anthropic has positioned itself as a safety-first alternative in the rapidly evolving AI sector. Its approach emphasizes rigorous testing, controlled deployment, and explicit limits on potentially harmful uses. That identity has helped it attract investors and enterprise clients seeking dependable guardrails — but it has also put the firm on a collision course with agencies that demand maximum operational flexibility.
Congressional and Industry Reactions
The standoff has reverberated across Washington. Advocacy organizations such as the Alliance for Secure AI, Common Cause, and Young Americans for Liberty have called on Congress to review the dispute, expressing concern that federal pressure on private companies could exceed legal boundaries.
Several lawmakers, including Senators Mark Warner and Chris Coons, have previously voiced unease about the rapid integration of AI into military systems without clear regulatory frameworks.
Legal scholars also note that designating Anthropic a supply chain risk — a label typically applied to foreign entities considered national security threats — would be an unusual move against a domestic AI company. Combining that with potential use of the Defense Production Act adds another layer of complexity, raising questions about how far executive authority can extend in compelling software companies to modify usage policies.
Competitive Pressures in the AI Race
The confrontation comes amid intense global competition for leadership in artificial intelligence. The U.S. government views advanced AI as essential to maintaining strategic superiority. Other major technology players, including OpenAI, Google, and xAI, have reportedly shown greater willingness to align their military-use policies with Pentagon expectations.
That dynamic places Anthropic in a delicate position. Standing firm could reinforce its safety-first brand but risk significant federal business. Conceding could secure continued contracts while potentially weakening the ethical commitments that distinguish it in a crowded AI marketplace.
