Secretary of Defense Pete Hegseth complains that it’s “woke AI” that SF’s own Anthropic won’t give him AI-controlled weapons and massive domestic surveillance, and is now threatening to blacklist the company.

If you don’t follow the AI wars much, there are four AI companies that could get involved with an actual war: The Pentagon has various AI contracts of $200 million or more with Google, OpenAI, Anthropic, and xAI. All four of those AI companies are based in the Bay Area, and Anthropic was formed by former OpenAI employees who felt they could build a more ethical AI company. Their coding and problem-solving product Claude is widely considered to be the best in the business.

But Anthropic’s ‘more ethical AI company’ philosophy has recently gotten them in trouble with the Pentagon, because the company refuses to allow their AI products to control weapons, or to conduct mass surveillance of US citizens. That may sound reasonable to most of us, but Defense Secretary Pete Hegseth complains that this is “woke AI” (whatever that even means?), and is in a pissing match with Anthropic to make the company concede to the Pentagon’s demands.

CNN reports that Hegseth met face-to-face on Tuesday with Anthropic CEO Dario Amodei. But it does not sound like Anthropic budged on keeping their guardrails in place.  

“A source familiar with the Tuesday meeting says the Pentagon plans to terminate Anthropic’s contract by Friday if the company does not agree to its terms,” according to CNN. “A Pentagon official told CNN the company has until 5:01pm on Friday to ‘get on board or not.’”

The Pentagon probably holds more cards in this conflict. They could invoke the Defense Production Act, a war powers-type act in which they can effectively force companies to do things if their products are considered essential. They could also declare Anthropic a “supply chain risk,” which would effectively blacklist them from doing business with the government and the many US companies who hold government contracts.

But it’s kind of a contradiction to blacklist a company just because their product is seen as essential.

“I would assume we don’t want to utilize the technology that is the supply chain risk, right? So I don’t know how you square that,” former Justice Department official Katie Sweeten told CNN. “What it sounds like is that the supply chain risk may not be a legitimate claim, but more punitive because [Anthropic is] not acquiescing.”

It’s all a fascinating, and kind of terrifying barometer of where we are with AI, particularly with an ignorant and bullying Trump administration in charge of the country. But it’s also kind of funny that despite all of the money that Elon Musk spent to elect Trump, Musk’s company xAI is not being given the Anthropic and Claude contracts, because xAI’s products are considered substantially crappier and less reliable.

Related: Anthropic's Pentagon Contract In Jeopardy Over Questions About AI Spying [SFist]

Image: WASHINGTON, DC - DECEMBER 16: Secretary of War Pete Hegseth moves from the Senate to the House as he attends closed door meetings with lawmakers on Capitol Hill on December 16, 2025 in Washington, DC. Hegseth and Secretary of State Marco Rubio are headed back to Capitol Hill to speak with lawmakers as questions mount about strikes carried out by the U.S. military on suspected drug boats out of Venezuela ordered by the Trump Administration. (Photo by Andrew Harnik/Getty Images)