Anthropic faces a critical deadline this Friday evening: either grant the U.S. military unfettered access to its advanced AI model or prepare for significant repercussions, Axios reports.
During a Tuesday morning meeting, Defense Secretary Pete Hegseth informed Anthropic CEO Dario Amodei of the Pentagon's two potential courses of action. The military could either label Anthropic a “supply chain risk,” a categorization typically reserved for hostile foreign entities, or invoke the Defense Production Act (DPA) to compel the company to customize its AI model for military applications.
The DPA empowers the president to mandate that companies prioritize or increase production in support of national defense initiatives. This act was notably utilized during the COVID-19 pandemic to require companies such as General Motors and 3M to manufacture essential ventilators and masks.
Anthropic has consistently articulated its firm stance against the use of its technology for mass surveillance of American citizens or for the development of fully autonomous weapons, steadfastly refusing to budge on these core principles.
Conversely, Pentagon officials contend that the military's deployment of technology must be governed by U.S. law and constitutional boundaries, rather than by the usage policies dictated by private contractors.
Employing the DPA in a disagreement concerning AI safeguards would represent a considerable expansion of the law's contemporary application. Furthermore, it would underscore an escalating pattern of executive branch instability that has intensified in recent years, as noted by Dean Ball, a senior fellow at the Foundation for American Innovation and former senior policy advisor on AI in Trump’s White House.
"It would essentially be the government declaring, ‘If you disagree with us politically, we’re going to attempt to put you out of business,’" stated Dean Ball, senior fellow at the Foundation for American Innovation and former senior policy advisor on AI in Trump’s White House.
This contention unfolds amid a landscape of ideological friction, with certain administration figures, including AI czar David Sacks, openly disparaging Anthropic’s safety protocols as "woke."
Ball further commented, "Any reasonable, responsible investor or corporate manager will observe this situation and conclude that the U.S. is no longer a stable environment for conducting business. This action strikes at the fundamental essence of what positions America as a crucial center for global commerce. We have consistently maintained a stable and predictable legal system."
This escalating standoff appears to be a high-stakes game of chicken, with Anthropic seemingly unwilling to yield. Reuters indicates that the company has no intention of relaxing its current usage restrictions.
Multiple reports highlight that Anthropic stands as the sole frontier AI laboratory possessing classified access with the Department of Defense. While the DOD currently lacks a readily available backup option, the Pentagon has reportedly secured an agreement to integrate xAI’s Grok into its classified systems.
Ball suggested that this absence of redundancy could be a driving factor behind the Pentagon’s assertive stance.
"Should Anthropic terminate the contract tomorrow, it would pose a significant challenge for the DOD," he informed TechCrunch, observing that the agency appears to be falling short of a National Security Memorandum issued by the late Biden administration, which instructs federal agencies to avoid reliance on a singular classified-ready frontier AI system.
He elaborated, "The DOD possesses no backups. This represents a single-vendor dependency. They cannot rectify that situation instantaneously."
TechCrunch has sought comments from both Anthropic and the DOD regarding these developments.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.