Pentagon Calls AI Firm a Security Risk, Then Uses It in Iran Strikes

"Too Dangerous to Trust, Too Valuable to Drop"

Defense Secretary Pete Hegseth labeled Anthropic a "Supply-Chain Risk to National Security" on Friday. Days later, U.S. forces used the company's Claude AI to support military strikes in Iran.

The contradiction is stark: If Anthropic truly threatens national security, why give it a six-month phase-out period? "OK, wait a minute, they're a really dangerous player for U.S. national security, so you're going to use them for another six months? Huh?" asked Herbert Lin, a Stanford researcher.

The Real Dispute: Control vs. Ethics

The breakdown between Anthropic and the Department of Defense centered on usage boundaries. The Pentagon wanted unfettered access to Claude models for "all lawful purposes." Anthropic pushed back, seeking assurances its technology wouldn't power fully autonomous weapons or domestic mass surveillance.

Until recently, Anthropic was the only AI company approved to deploy models across the Pentagon's classified networks. While OpenAI and Elon Musk's xAI have received clearance, transitioning systems takes significant time and resources.

Politics Masquerading as Policy

Experts see this as more political theater than genuine security concern. The Trump administration hasn't outlined specific technical threats or security breaches. Instead, critics cite Anthropic's "arrogance" and resistance to military demands.

CEO Dario Amodei took a different approach than other tech executives, avoiding early courtship of the Trump administration. Conservatives have repeatedly accused the company of pushing "woke AI." Michael Horowitz from the Council on Foreign Relations calls it "a dispute that is about politics and personalities."

The Market Reality Check

Despite the public spat, the Pentagon's continued use of Claude in Iran operations sends a clear signal about the technology's value. "There's no clearer signal" of how much the military values this capability, Horowitz noted.

Defense contractors face a practical dilemma. Some are switching vendors "out of an abundance of caution," while others wait for formal guidance. The uncertainty creates costs and efficiency losses across the defense industrial base.

Supply Chain Risk Without the Paperwork

Anthropic hasn't received an official supply chain risk designation beyond social media posts. The company has threatened to challenge any formal designation in court. Samir Jain from the Center for Democracy and Technology argues the legal requirements for such designation likely haven't been met.

Yet the market is already responding. Venture capitalists report portfolio companies switching providers, and several defense executives told CNBC they're moving away from Anthropic's models.