Anthropic CEO Dario Amodei acknowledged that the U.S. government classified his company as a supply chain risk on Thursday and stated it has “no option” but to contest the classification legally.
The startup has had disagreements with the Department of Defense regarding the permissible applications of its artificial intelligence models, recognized as Claude, and was informed late last week, through social media updates, that it faced a ban from government contracts.
Anthropic requested confirmation that its technology would not be utilized for fully autonomous weapons or comprehensive domestic surveillance, however, the DOD wanted Anthropic to allow unrestricted access to Claude for all legal uses.
“As we mentioned last Friday, we do not believe, and have never believed, it is the role of Anthropic or any private entity to participate in operational decision-making—that responsibility lies with the military,” Amodei stated. “Our primary concerns have been our objections to fully autonomous weapons and extensive domestic surveillance, which pertain to critical usage areas, not operational decision-making.”
Anthropic is the sole American firm ever publicly identified as a supply chain risk, and this official classification will compel defense contractors to verify they do not incorporate the company’s models in their dealings with the Pentagon. This designation has generally been reserved for entities associated with foreign adversaries, such as the Chinese tech firm Huawei.
Questions linger regarding whether defense contractors can use Anthropic’s technology for endeavors outside their military engagements. Amodei noted in his statement that the classification “does not (and cannot) restrict the use of Claude or business engagements with Anthropic if those are unrelated to their specific Department of War contracts.”
Microsoft, which revealed intentions to invest up to $5 billion in Anthropic in November, expressed in a statement that its attorneys “examined the classification” and concluded that Anthropic’s products can remain accessible to clients aside from the DOD.
Anthropic entered into a $200 million agreement with the DOD in July, becoming the first AI laboratory to integrate its models into mission workflows on classified networks. However, as negotiations between both parties stalled, competitors OpenAI and Elon Musk’s xAI also reached agreements to deploy their models for classified uses.
OpenAI CEO Sam Altman disclosed his company’s arrangement with the DOD just hours after Anthropic was blacklisted on Friday. He stated in a message on X that the agency showed a “strong commitment to safety and a desire to collaborate for the best possible outcome.”
Anthropic’s engagement with the Trump administration has become increasingly strained in recent months, and Amodei expressed regret for a sensitive internal memo that was leaked to the media on Wednesday.
Amodei reportedly informed staff that the administration disapproves of Anthropic because it has not made donations or given “dictator-style accolades to Trump,” according to a report from The Information.
He clarified that the memo was drafted on Friday following a “challenging day for the company” and does not represent his “thoughtful or measured views.” Amodei mentioned that it is an “outdated assessment of the current status.”
“Anthropic did not divulge this post nor direct anyone else to do so—it is not in our interest to escalate this matter,” Amodei asserted.
WATCH: Sec. Pete Hegseth instructs Pentagon to label Anthropic as a supply-chain risk to national security
