Pentagon vs Anthropic: U.S. AI Use in Military Sparks Legal Battle

A high-stakes clash over how the U.S. military should use artificial intelligence has spilled into the open this week as Defense Secretary Pete Hegseth abruptly ended the Pentagon’s work with Anthropic and other government partners, applying a law designed to counter foreign supply chain threats to a domestic tech firm. The move effectively labels the San Francisco–based startup as a risk to the defense supply chain and could end a contract worth up to $200 million.

Anthropic criticized the action as legally unsound and vowed to challenge it in court, calling the move unprecedented for an American company. Its leaders have warned that a portion of its AI products could be used for mass surveillance or autonomous weapons, a stance that helped fuel a broader debate about how safe AI should be in military applications. The company’s supporters argue that surrendering safety controls to a government mandate could chill innovation and create a dangerous precedent.

The clash also unsettles the broader tech landscape. OpenAI, once a rival to Anthropic, has stepped into the void by offering its own tools to the Pentagon after Anthropic objected to some terms of the government’s push. That turn of events underscores how quickly power dynamics can shift in Big Tech amid security concerns, and it highlights the personal rivalries that have shaped AI policy, including tensions between OpenAI’s Sam Altman and Anthropic’s Dario Amodei.

Beyond personalities, the episode raises questions about the Defense Production Act, a Cold War–era authority typically used to safeguard physical goods and information flows rather than to alter AI safety features. Reports describe the Pentagon’s willingness to invoke that act to compel access to Claude, Anthropic’s flagship model, illustrating how policymakers are balancing national security fears with innovation incentives.

As the legal and political drama unfolds, one clear consequence could be a reordering of who gets to set guardrails for AI in defense. If Anthropic fights this in court and sustains its position, the industry could gain ground in shaping how and when military AI is deployed, with OpenAI and other players watching closely for how far policymakers will go in policing domestic AI tools.

#AnthropicVsPentagon #AIinWarfare #DefenseTech