Pentagon Faces Lawsuit from Anthropic Over Supply Chain Risk
Anthropic PBC has pledged to challenge in court a Pentagon ruling that categorizes the company as a threat to the US supply chain, a designation typically applied to foreign adversaries. This move intensifies the conflict with the Trump administration regarding the regulation of artificial intelligence safeguards. “We do not believe this action is legally sound, and we see no choice but to challenge it in court,” Dario Amodei stated. Amodei stated in his post that defence officials had informed the company on Wednesday regarding the supply-chain designation. The decision jeopardizes the firm’s $200 million contract to supply the Pentagon with classified AI tools, and may prevent Anthropic from collaborating with other companies on their defense projects. The decision marked the conclusion of weeks filled with escalating tensions during negotiations between Amodei and government officials regarding the US military’s access to Anthropic’s technology. Discussions collapsed last week when the company sought guarantees that its AI would not be utilized for mass surveillance of Americans or for the deployment of autonomous weapons. This led Defence Secretary Pete Hegseth to warn the company about the potential supply-chain risk designation.
Despite Anthropic’s intention to contest the decision, Amodei stated that the statute referenced — section 3252 of the US law governing the armed forces — is sufficiently specific to prevent it from impacting other Anthropic operations that do not pertain to particular Pentagon contracts. This provides a measure of comfort for customers and investors concerned that the company might be unable to engage in business with entities associated with the Pentagon. A spokesperson for Microsoft Corp. stated on Thursday that the company has determined it can proceed with collaborations with Anthropic on non-defense initiatives. Nonetheless, the designation requires the company to cease its collaboration with Palantir Technologies Inc., another contractor serving the military. Palantir’s integration of Anthropic’s Claude into the digital mission control platform, Maven Smart System, is noteworthy, particularly as it has been utilized in the US military’s campaign in Iran.
The Pentagon’s decision poses a risk of hindering wider initiatives aimed at speeding up the integration of AI within the US military. Until recently, Anthropic was the sole provider of an AI system capable of functioning within the Pentagon’s classified cloud, and its Claude Gov tool has emerged as a preferred choice among defense personnel due to its user-friendly design. “It’s a good capability,” and removing it is “going to be painful for all involved,” stated Lauren Kahn. On Thursday, a defence official stated that the determination would be “effective immediately.” However, a source familiar with the situation revealed that Anthropic’s Claude AI tools continue to be actively utilized by the US military in operations in Iran. In his warning to the firm last Friday, Hegseth outlined a six-month transition period to shift AI work to other providers. Hegseth has communicated his decision to Congress through letters addressed to leading Republicans and Democrats on the House and Senate committees for armed services, appropriations, and intelligence, as per correspondence. “This determination is based in part on a risk analysis by the DoW and input from senior DoW personnel that the Covered Entity’s restrictions on the use of its products and services introduces national security risks to the DoW’s supply chain,” Hegseth wrote, referring to Anthropic and using an acronym for the Department of War, the name he now favors for the Department of Defense.
Anthropic is preparing for potential job losses at civilian agencies, in light of President Donald Trump’s recent request for the federal government to terminate the firm’s contract. The Treasury Department and General Services Administration have already declared their plans to cease business with the company. In his post Thursday, Amodei stated that he had been engaged in ongoing discussions with the Pentagon over the past several days, which he regarded as “productive,” regarding how to address the company’s concerns on AI safeguards. He stated that Anthropic intended to continue supplying its products to the military for as long as it is allowed. Emil Michael, the under secretary of defence for research and engineering who had been negotiating over the past several weeks with Amodei, stated in an X post late Thursday that there were no further discussions. “I want to end all speculation: there is no active @DeptofWar negotiation with @AnthropicAI,” he wrote. Currently valued at $380 billion, Anthropic is poised to achieve an annual revenue of nearly $20 billion, more than doubling its run rate from late last year. The Pentagon dispute, however, has complicated the outlook for the company. “Applying a supply-chain risk label to an American company would be unprecedented and go beyond the scope of the law,” according to Charlie Bullock.
“This is not an authority that’s meant for destroying large American companies that have a contractual disagreement with the United States government,” he stated. “It’s an authority that’s meant for addressing spying by Chinese companies and stuff like that.” Recognizing the challenging dynamics of the discussions, Amodei expressed regret for remarks made in an internal memo that came to light on Wednesday. In the memo, as reported, Amodei accused arch-rival OpenAI of acting opportunistically and sacrificing safeguards in a Pentagon deal that CEO Sam Altman announced just hours after Trump and Hegseth ordered an end to US government ties to Anthropic. In the memo, Amodei stated that he believes the true reason the administration has a negative view of Anthropic is due to the company’s lack of donations to Trump, failure to support his AI policies, or provide him with “dictator-style praise.” On Thursday, Amodei stated that “Anthropic did not leak this post nor direct anyone else to do so.” He said “It is not in our interest to escalate,” adding that it was written hours after posts from Trump and Hegseth, as well as OpenAI’s announcement. “It was a difficult day for the company, and I apologise for the tone of the post.”
