In a first, Trump Sanctions American Firm, Anthropic Inc

A contract dispute between the U.S. government and an AI startup spilled into public view on Friday after U.S. President Donald Trump ordered every federal agency to stop using technology developed by Anthropic, citing national security concerns and accusing the company of trying to dictate military policy.

In a post written in all caps on Truth Social, Trump said, “THE UNITED STATES OF AMERICA WILL NEVER ALLOW A RADICAL LEFT, WOKE COMPANY TO DICTATE HOW OUR GREAT MILITARY FIGHTS AND WINS WARS!” He added that Anthropic had made a “DISASTROUS MISTAKE” by attempting to “STRONG-ARM” the Department of War and announced an immediate halt to its technology across federal agencies.

While most agencies must stop using Anthropic’s tools right away, the Pentagon has been granted a six-month period to phase out systems already embedded in military and intelligence workflows.

At the center of the standoff is Anthropic’s AI model, Claude, which has been deployed on classified U.S. government networks since June 2024. The company sought what it described as “narrow assurances” that its technology would not be used for mass domestic surveillance or integrated into fully autonomous weapons capable of lethal action without human oversight.

Pentagon officials said they had no intention of using the AI for those purposes and would operate within the law. However, they insisted on access “without any limitations.”

Anthropic’s chief executive, Dario Amodei, said the company “cannot in good conscience accede” to demands that could allow safeguards to be overridden. In a statement released after the White House order, the company said: “No amount of intimidation or punishment from the Department of War will change our position on mass domestic surveillance or fully autonomous weapons. We will challenge any supply chain risk designation in court.”

Anthropic argued that current frontier AI models are not reliable enough for deployment in fully autonomous weapons systems and warned that mass surveillance of Americans would violate fundamental rights.

Defense Secretary Pete Hegseth escalated the matter by designating Anthropic a “supply chain risk to national security,” a label historically applied to foreign adversaries. He said the Pentagon “must have full, unrestricted access to Anthropic's models for every LAWFUL purpose in defense of the Republic.”

Pentagon spokesman Sean Parnell claimed Anthropic’s stance was “jeopardizing critical military operations and potentially putting our warfighters at risk.”

The dispute has reverberated through Silicon Valley. Elon Musk sided with the administration, writing on X that “Anthropic hates Western Civilization.” His company’s Grok AI system is expected to gain access to classified military networks.

In contrast, Sam Altman, chief executive of OpenAI, expressed support for Anthropic’s position on limits. “For all the differences I have with Anthropic, I mostly trust them as a company, and I think they really do care about safety,” Altman was quoted as saying by CNBC. He later said OpenAI had reached its own agreement to deploy models within the Department of War’s classified systems with safety guardrails reflected in the contract.

Anthropic said that any formal designation under 10 USC 3252 would apply only to Department of War contract work and would not affect commercial customers or individual users. Legal experts cited in U.S. media reports questioned whether the supply chain risk label was intended for disputes involving American companies in active contract negotiations.

Claude is currently used by U.S. intelligence agencies, including the CIA and NSA, for analysis tasks. The six-month phase-out could require agencies to transition to alternative AI systems while legal and policy challenges unfold.