The U.S. Department of Defense has issued a stark ultimatum to Anthropic, the sole provider of artificial intelligence models currently integrated into classified military systems. By Friday at 5:01 p.m., Anthropic must comply with Pentagon demands or face the activation of the Defense Production Act (DPA). The DPA would compel Anthropic to allow military use of its AI model, while simultaneously designating the company as a supply chain risk – a move that could jeopardize lucrative government contracts.
The Contradiction: This dual threat – preventing and forcing government access – highlights the Pentagon’s growing frustration with Anthropic’s resistance. The urgency underscores just how critical Anthropic’s technology has become for military operations.
Defense Secretary Pete Hegseth personally summoned Anthropic CEO Dario Amodei to the Pentagon on Tuesday for a meeting, where the ultimatum was delivered. Sources say the discussion was outwardly polite, yet firm: agree to the terms, or face consequences.
Why This Matters: The Pentagon’s aggressive posture is unusual, especially given that the DPA and supply chain risk designations are typically reserved for genuine national security emergencies. By weaponizing these tools as leverage in a business dispute, the Pentagon risks undermining their credibility and diluting future enforcement efforts. As Jessica Tillipman, an associate dean at George Washington University Law School, points out, “The bigger issue here is that it waters down these designations… transforming national security tools into a point of leverage for business.”
The move reflects a broader trend: governments worldwide are grappling with how to regulate AI, particularly when it comes to defense applications. The Pentagon’s tactics signal a willingness to use extreme pressure to secure access to critical technologies, raising questions about the balance between national security and corporate autonomy.
The Outcome: Anthropic must decide whether to yield to the Pentagon’s demands or risk losing access to valuable government contracts. The outcome will likely set a precedent for future interactions between the military and AI developers, and will shape how quickly the armed forces adopt advanced AI systems.
