Secretary of War Pete Hegseth has given Anthropic CEO Dario Amodei until Friday evening to grant the military unrestricted access to its AI model or face significant penalties, according to sources familiar with the matter.
Axios reports that the ultimatum was delivered during a tense meeting on Tuesday, where Hegseth informed Amodei that the Pentagon would either terminate its relationship with Anthropic and designate the company as a supply chain risk, or invoke the Defense Production Act to compel the company to customize its model for military requirements without safeguards.
The confrontation represents an escalation in the ongoing dispute between the Pentagon and Anthropic over AI safety protocols. Defense officials are caught between wanting to penalize the company for maintaining certain usage restrictions and recognizing their dependence on Claude, Anthropic’s industry-leading AI model. A Defense official acknowledged this tension, stating that the only reason discussions continue is because the military needs Anthropic’s capabilities immediately, noting that the company’s technology is exceptionally advanced.
Anthropic has indicated willingness to modify its usage policies for Pentagon applications, but maintains firm boundaries against allowing its model to be deployed for mass surveillance of American citizens or for developing autonomous weapons systems that can engage targets without human oversight. Currently, Claude stands as the sole AI model authorized for the military’s most classified operations.
According to a senior official, the atmosphere during Tuesday’s meeting was distinctly unfriendly. However, another source characterized the discussion as remaining cordial throughout, with no raised voices from either party, and noted that Hegseth actually commended Claude’s performance to Amodei.
During the meeting, Hegseth made clear that the Pentagon would not permit any private company to dictate operational decision-making terms or raise objections to specific use cases. He specifically referenced Pentagon claims that Anthropic had expressed concerns to its partner company Palantir regarding Claude’s deployment during a recent military operation targeting Venezuelan leader Nicolas Maduro.
Amodei firmly denied that Anthropic had raised any such concerns or discussed the matter with Palantir beyond routine operational communications. He emphasized that the company’s established boundaries have never interfered with Pentagon operations or created problems for personnel in the field.
Following the meeting, Anthropic maintained a diplomatic stance. A company spokesperson said Amodei expressed gratitude for the Department’s work and thanked Hegseth for his service. The spokesperson added that productive discussions continued regarding usage policy adjustments to ensure Anthropic can support national security missions within the boundaries of what their models can reliably and responsibly accomplish.
The Defense Production Act, which Hegseth threatened to invoke, grants presidential authority to require private companies to accept and prioritize specific contracts deemed necessary for national defense. The law saw application during the COVID-19 pandemic for increasing vaccine and ventilator production. However, using it in such an explicitly confrontational manner would be highly unusual. The Pentagon’s intention would be forcing Anthropic to adapt its model for military purposes without any protective safeguards.
The Pentagon is simultaneously considering contract termination and designating Anthropic as a supply chain risk, which would require numerous other Pentagon contractors to verify that Claude is not integrated into their operations.
Severing ties with Anthropic presents complications, as the Pentagon would need a ready replacement for Claude, currently the exclusive model deployed in classified systems. Claude’s military applications extend from its use in the Venezuela operation through Anthropic’s Palantir partnership to various administrative functions throughout the military.
The future of autonomous weapons and how AI factors into national security is a major topic of the upcoming book by Breitbart News Social Media Director Wynton Hall, Code Red: The Left, the Right, China, and the Race to Control AI:
In Code Red, Breitbart social media director Wynton Hall exposes where that power hides, how it operates, how conservatives can navigate the AI political battlescape, avert its landmines, and turn peril into promise. AI decides what you see and what gets censored. It’s quietly rewiring our whole way of life. Jobs. Schools. Family. Church. Even national security. All of it will shock-test our civic order.
Inside Code Red, you will discover:
- Why AI is wired for woke indoctrination—and how to resist it.
- How elites plan to weaponize AI job losses to push dependency.
- How America can beat China without becoming China.
- How to prepare your kids for the blinding speed of AI disruption.
- The new national security threats AI unleashes—and how we defend against them.
- Why “AI girlfriends” are luring millions—and what it will take to preserve authentic human connection.
- How AI will test faith and meaning—and why spiritual renewal may be its most surprising outcome.
Urgent, deeply researched, and written with page-turner elegance, Code Red is the conservative battle plan for the AI era. Either we wake up and fight back, or we lose everything that made America free.
Read more at Axios here.
Lucas Nolan is a reporter for Breitbart News covering issues of free speech and online censorship.
Read the full article here