Anthropic lose-lose situation in Pentagon demand for AI policy change
Anthropic enters Friday in a winless situation.
The AI startup has until 5:01 p.m. ET to decide whether to allow the Department of Defense to use its models without restrictions in all legitimate use cases. Otherwise, Defense Secretary Pete Hegseth threatened to label the company a “supply chain risk” or force the company into compliance by invoking the Defense Production Act.
Antropik signed 200 million dollars It signed a contract with the Department of Defense in July, becoming the first AI lab to integrate its models into mission workflows in covert networks. The company is negotiating the terms of its deal with the agency and has sought assurances that its technology will not be used for fully autonomous weapons or domestic mass surveillance of Americans.
“We believe that in a narrow set of situations, AI can undermine democratic values rather than defend them,” Anthropic CEO Dario Amodei, who co-founded the company in 2021, said in a statement Thursday. “Some uses are also beyond the limits of what today’s technology can do safely and reliably.”
The Ministry of Defense refused to back down, and negotiations descended into a stalemate that became the highest-profile test to date of Anthropic’s stated values. The company has spent years carefully building its reputation as a champion of safe and responsible AI deployment, positioning itself in contrast to OpenAI, where Amodei worked before leaving to start Anthropic.
But Anthropic also faces intense pressure to justify its massive $380 billion valuation, backed by major institutional and strategic investors, while racing to stay at the cutting edge of model development and fend off competition from OpenAI and other rivals. Google and Elon Musk’s xAI. All three of these companies’ models are used by the Department of Defense.
Giving in to DoD demands could damage Anthropic’s reputation and alienate employees and customers. But if Anthropic refuses to provide the military with unrestricted access to its models, it could lose meaningful revenue in the short term and miss out on potential future opportunities with other companies doing business with the government.
“There are no winners in this,” Lauren Kahn, senior research analyst at the Georgetown Center for Security and Emerging Technology, told CNBC in an interview. “It leaves a sour taste in everyone’s mouth.”
The Pentagon’s chief spokesman, Sean Parnell, said Thursday that the Defense Department has “no interest” in using artificial intelligence for fully autonomous weapons or conducting mass surveillance on Americans, calling it illegal. He said he wanted the agency to agree to allow Anthropic’s models to be used “for all lawful purposes.”
“This is a simple, common-sense request that will prevent Anthropic from jeopardizing critical military operations and potentially putting our warfighters at risk,” Parnell said. he wrote. a post Thursday at X. “We will not allow ANY company to dictate the terms of how we make operational decisions.”
One separate post On Thursday, the undersecretary of defense for research and engineering and former Uber The admin wrote that Amodei was “a liar and has a God complex.” He accused Amodei of wanting “nothing more than trying to personally control the US Army.”
In a meeting with Amodei earlier this week, Hegseth set Anthropic’s deadline for Friday and warned that the penalty for disagreeing could be severe. He said Anthropic could be labeled a “supply chain risk”; This is a specification for companies in countries that are often viewed as competitors. This label will force DoD vendors and contractors to certify that they are not using Anthropic’s models.
Amodei said his company won’t be intimidated.
“These threats do not change our position: we cannot in good conscience accept their demands,” he said in a statement on Thursday. he wrote.
‘The juice is not worth squeezing’
The escalating conflict is one that other AI labs, industry experts and government contractors are watching closely. Kahn warned that the government could alienate tech companies with promising products if they conclude that “the juice is not worth the squeeze.”
“I’m really, really, really worried about private companies saying, ‘It’s not worth spending time moving forward with the defense sector,'” Kahn said, adding that “the ones who will really suffer will be the war fighters.”
OpenAI CEO Sam Altman told CNBC on Friday that he “personally doesn’t think the Pentagon should threaten the DPA against these companies.” He said he thinks it’s important for companies to choose to work with the department as long as it complies with legal protections and “a few red lines” the field shares with Anthropic.
“Despite all the differences I have with Anthropic, I mostly trust them as a company and I think they really care about safety and I’m happy they support our war warriors,” Altman said in an interview. “I’m not sure where this is going.”
Scores of employees from Anthropic and elsewhere in the industry have taken to social media in recent days to express their support for the company.
Josh McGrath, technical staff at OpenAI, wrote: a post He was “at a loss for words about what’s going on” at X on Tuesday, he said.
According to the website, more than 330 employees from Google and OpenAI also signed an open letter from the ministry titled “We Will Not Be Divided”, which aims to create “common understanding and solidarity in the face of this oppression”.
“We hope that our leaders will put their differences aside and stand together to reject the War Department’s current requests for permission to use our models for domestic mass surveillance and autonomous killing of people without human supervision,” the letter says.
United States Secretary of War Pete Hegseth speaks during a visit to Sierra Space in Louisville, Colorado, Monday, February 23, 2026.
Aaron Ontiveroz | Denver Post | Getty Images
According to Anthropic, this is the latest conflict with the Trump administration.
Venture capitalist David Sacks, who serves as the White House AI and crypto czar, previously accused Anthropic of promoting “woke AI” and “executing a complex regulatory capture strategy based on fear mongering” over its stance on regulation after a company executive wrote an article titled “Technological Optimism and Appropriate Fear” in October.
Amodei has also largely avoided rubbing elbows with President Donald Trump, unlike other industry executives, including Altman. Apple CEO Tim Cook and Google CEO Sundar Pichai. Amodei notably did not attend Trump’s inauguration last year.
In January, Hegseth published a memo titled “Accelerating America’s Military AI Dominance.” In the document, he wrote that the Department of Defense should not use AI models that “involve ideological ‘tuning'” and said the department “should also use models that are free of use policy restrictions that could limit lawful military applications.”
While Amodei remains committed to using his company’s models safely, he said Thursday that Anthropic’s “strong preference” is to continue working with the Department of Defense and supporting U.S. national security.
“Should the Department choose to exit Anthropic, we will work to ensure a smooth transition to another provider, avoiding any disruption to ongoing military planning, operations or other critical missions,” Amodei wrote.
— CNBC’s Kate Rooney contributed to this report
WRISTWATCH: Antropik rejects Pentagon’s ‘final offer’ in fight for AI protection
