As of Friday, Anthropic appeared to be in a no-win position. The company faced a deadline of 5:01 PM Eastern Time to decide whether to grant the U.S. Department of Defense unrestricted use of its AI models in all legal scenarios. Refusal could lead the Pentagon, under Secretary Hegseth, to designate the company as a "supply chain risk" or invoke the Defense Production Act to compel compliance.

In July 2025, Anthropic signed a $200 million contract with the Department of Defense, becoming the first AI lab to integrate its models into classified network mission workflows. The company had been negotiating terms, seeking assurances that its technology would not be used for fully autonomous weapons or domestic mass surveillance of American citizens.

In a statement on Thursday, Anthropic CEO Dario Amodei wrote, "In a small number of specific contexts, we believe AI can undermine democratic values rather than defend them. Some applications also fall entirely outside the scope of what current technology can operate safely and reliably."

With the Pentagon refusing to concede, negotiations reached a stalemate, presenting the most significant public test yet of Anthropic's professed values. For years, the company has cultivated an image as an advocate for the safe and responsible deployment of AI, contrasting with the approach of OpenAI, where Amodei previously worked.

Simultaneously, Anthropic is under immense pressure to justify its $380 billion valuation—backed by large institutions and strategic investors—while maintaining a lead in model development against competitors like OpenAI, Alphabet, and Elon Musk's xAI, all of which have models already adopted by the U.S. Defense Department.

CapITulating to the Pentagon's demands could damage Anthropic's reputation and alienate employees and clients. However, refusing to grant the military unrestricted access to its models could result in significant lost revenue in the short term and potentially exclude the company from future government partnerships.

Lauren Kahn, a senior research analyst at Georgetown University's Center for Security and Emerging Technology, stated in an interview, "There are no winners here. This just leaves everyone feeling uncomfortable."

On Thursday, Pentagon press secretary Sean Parnell stated that the Defense Department has "no intention" of using AI for fully autonomous weapons or for mass surveillance of U.S. citizens, noting such actions are already illegal. He said the Pentagon wants Anthropic to agree that its models can be used for "all lawful purposes."

Parnell posted on X, "This is a simple, common-sense request to prevent Anthropic from jeopardizing critical military operations or putting our warfighters at risk. We will not allow any company to dictate how we make combat decisions."

Emil Michael, the Undersecretary of Defense for Research and Engineering and a former Uber executive, also posted on X, calling Amodei a "liar with a god complex" and accusing him of "wanting personal control over the U.S. military."

Earlier in the week, Secretary Hegseth set the Friday deadline during a meeting with Amodei, warning that non-compliance would be met with severe penalties. He indicated Anthropic could be designated a "supply chain risk"—a label typically applied to firms from adversary nations. Such a designation would require Defense Department suppliers and contractors to prove they do not use Anthropic's models.

Amodei stated the company would not be intimidated.

He wrote in Thursday's statement, "These threats will not change our position: we cannot in good conscience agree to their demands."

**A Losing Proposition**

The escalating conflict is being closely watched by other AI labs, industry experts, and government contractors. Kahn warned that if companies decide the downsides outweigh the benefits, the government risks locking out tech firms with superior products.

Kahn said, "I am genuinely, sincerely concerned that private companies will say, 'It's not worth it to work with the Defense Department going forward.' And the ones who lose out are the warfighters."

In a CNBC interview on Friday, OpenAI CEO Sam Altman said he "personally believes the Pentagon should not be threatening these companies with the Defense Production Act." He argued that as long as companies adhere to legal safeguards and the "few red lines" shared by the industry and Anthropic, it is important for firms to choose voluntarily to work with the Defense Department.

Altman said in the interview, "While I have many disagreements with Anthropic, I generally trust the company and believe they take safety seriously. I'm glad they have been supporting our warfighters. It's unclear how this will play out."

Recently, numerous employees from Anthropic and other companies in the industry have expressed support for the firm on social media.

OpenAI technician Josh McGrath posted on X on Tuesday, stating he was "speechless about what is happening."

According to a related website, over 330 employees from Alphabet and OpenAI have signed an open letter titled "We Will Not Be Divided," aiming to "build consensus and solidarity under pressure from the Defense Department."

The letter states, "We hope leadership will set aside differences, stand together, and continue to refuse the current demands from the War Department—namely, allowing the use of our models for domestic mass surveillance and for autonomous killing without human oversight."

**Context**

For Anthropic, this is just the latest conflict with the Trump administration.

David Sacks, the White House AI and Crypto Czar and a venture capitalist, previously accused Anthropic of promoting "woke AI" due to its regulatory stance, claiming the company employs "a sophisticated regulatory capture strategy based on fearmongering." This followed an article titled "Techno-Optimism and Moderate Fear" published by an Anthropic executive in October 2025.

Unlike other industry leaders such as Sam Altman, Apple CEO Tim Cook, and Alphabet CEO Sundar Pichai, Amodei has largely avoided engagement with President Trump. Notably, he did not attend Trump's inauguration last year.

In January, Secretary Hegseth released a memo titled "Accelerating U.S. Military AI Dominance." He wrote that the Defense Department must not adopt AI models that incorporate "ideological 'tuning'," and stated the department "must use models free from usage policies that restrict lawful military applications."

Despite Amodei's commitment to the company's principles on model safety, he stated on Thursday that Anthropic "strongly desires" to continue working with the Defense Department to support U.S. national security.

Amodei wrote, "If the Defense Department chooses to end its partnership with Anthropic, we will assist in a smooth transition to other providers to avoid any disruption to current military planning, operations, and other critical missions."