Anthropic CEO Stands Firm as Pentagon Deadline Looms

When Defense Secretary Pete Hegseth set the deadline, he made it clear: 5:01 p.m. Friday, or else. The ultimatum delivered to Anthropic this week represents one of the most direct confrontations yet between the U.S. military and an AI company over the boundaries of artificial intelligence use. With hours remaining before the Pentagon’s deadline, Anthropic CEO Dario Amodei has made his position unmistakable: he will not comply.

“We cannot in good conscience accede to their request.” — Dario Amodei, CEO, Anthropic

The Standoff

The dispute centers on two specific uses of AI that Anthropic refuses to enable: mass surveillance of American citizens and fully autonomous weapons systems operating without human oversight. The Pentagon believes it should have unrestricted access to Anthropic’s Claude AI model for all lawful military purposes. Amodei believes some lawful purposes should remain off-limits.

The disagreement has escalated rapidly. Earlier this week, Pentagon officials met with Anthropic and delivered a stark choice: agree to unrestricted military use by Friday’s deadline, or face consequences under the Defense Production Act. The Cold War-era law, last invoked during the COVID-19 pandemic, grants the federal government sweeping authority to compel private companies to prioritize national defense needs.

The supply chain threat represents an even more damaging potential outcome. Pentagon officials have threatened to label Anthropic a supply chain risk—a designation typically reserved for companies from adversary nations. Such a label could severely damage Anthropic’s ability to work with the U.S. government and harm its reputation across the technology sector.

“One labels us a security risk; the other labels Claude as essential to national security.” — Dario Amodei

The Ethical Line

Amodei’s objections are specific and grounded in the safety-focused philosophy that has defined Anthropic since its founding. The company was established in 2021 by former OpenAI employees who believed AI development should prioritize safety and ethical considerations alongside capability.

Mass domestic surveillance represents the first red line. Amodei has been unequivocal: “Using these systems for mass domestic surveillance is incompatible with democratic values.” The concern is not theoretical—AI systems capable of analyzing vast quantities of communications and behavioral data could enable monitoring at a scale previously unimaginable.

Fully autonomous weapons constitute the second boundary. Amodei argues that current AI systems are not reliable enough to be trusted with deadly force without human oversight. “Leading AI systems are not yet reliable to be trusted to power deadly weapons without a human in ultimate control,” he stated. “We will not knowingly provide a product that puts America’s warfighters and civilians at risk.”

The Pentagon’s position is equally clear: legality is the Pentagon’s responsibility as the end user, not the technology provider’s. A senior defense official pushed back on Anthropic’s concerns, insisting that the Department of Defense “has only given out lawful orders.”

The Competitive Pressure

Anthropic’s resistance comes amid intensifying competition for military AI contracts. The company was selected last year alongside OpenAI and Google to supply AI models for military applications under a $200 million agreement. At the time, the contract represented validation of Anthropic’s technology and a significant revenue opportunity.

The landscape has shifted. Pentagon officials confirmed this week that Elon Musk’s xAI has received clearance for classified use of its Grok system. OpenAI and Google are reportedly close to similar clearances. The message to Anthropic is implicit: alternatives exist, and the military will not be held hostage to a single vendor’s ethical constraints.

Yet Anthropic maintains a unique position. According to defense officials, the company currently operates the only frontier AI system with classified-ready status for military use. A forced transition to another provider would not be seamless—it would disrupt ongoing military planning, operations, and critical missions.

“Anthropic understands that the Department of War, not private companies, makes military decisions. However, in a narrow set of cases, we believe AI can undermine, rather than defend, democratic values.” — Dario Amodei

The Broader Implications

The confrontation raises fundamental questions about the relationship between private AI companies and government power. As AI capabilities advance, these systems are becoming essential infrastructure for national security. The question of who controls their use—and according to what principles—will only grow more urgent.

The Defense Production Act represents nuclear-option territory. Invoking it would set a precedent for direct government control over AI development and deployment. Such a move would likely face legal challenges and could reshape the landscape of public-private partnerships that have driven American technological leadership.

The supply chain designation would be equally consequential. Treating an American AI company as equivalent to firms from adversary nations would signal a dramatic shift in how the government views the technology sector. The implications would extend far beyond Anthropic to every company developing sensitive technologies.

What Comes Next

As the Friday deadline approaches, both sides have left room for negotiation while holding firm on core principles. Amodei has emphasized Anthropic’s desire to continue serving the military—with safeguards in place. He has also offered to facilitate a smooth transition to another provider if the Pentagon chooses to end the relationship.

The Pentagon, for its part, has not publicly committed to invoking the Defense Production Act or issuing the supply chain designation. The threats remain credible precisely because they have not been executed—giving Anthropic every incentive to reconsider its position before the deadline.

Whatever happens Friday evening, the confrontation has already exposed the fault lines in America’s AI strategy. The military needs cutting-edge capabilities. AI companies need ethical boundaries. Reconciling those needs will define the next chapter of both American technology and national security.

This article was reported by the ArtificialDaily editorial team. For more information, visit TechCrunch and The Hindu.