Trump states that the Anthropic deal with the Pentagon is "possible."

      On Tuesday, the US president informed CNBC that Anthropic is "shaping up" following a White House meeting last Friday where the company's CEO, Dario Amodei, discussed its Mythos AI model with Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent.

      The Pentagon's decision to blacklist Anthropic remains unresolved in court, as a federal appeals court and a district court in San Francisco have issued conflicting decisions. President Donald Trump stated in an interview on CNBC’s Squawk Box that a potential agreement allowing the use of Anthropic's AI models within the Department of Defense is "possible," describing the company as "shaping up."

      "They visited the White House a few days ago, and we had some very productive discussions with them, and I believe they are shaping up," Trump remarked. "They are quite intelligent, and I think they could be a significant asset." These comments represent a notable shift from Trump's stance in late February when he posted on Truth Social directing all federal agencies to "IMMEDIATELY CEASE all use of Anthropic’s technology" and stated his administration would "not do business with them again."

      Trump's comments followed a meeting on April 18, during which Anthropic CEO Dario Amodei met with Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent to discuss the company's new Mythos model—a cutting-edge AI system that Anthropic claims excels in cybersecurity tasks and has only been made available to a select few organizations. The White House characterized the conversation as "productive and constructive." Anthropic noted that Amodei had a "productive discussion" with government officials on how the company and the US government could "collaborate on key shared priorities like cybersecurity, America’s leadership in the AI sector, and AI safety."

      When asked about the meeting while at a Phoenix airport, Trump initially responded with "Who?" and claimed he was unaware Amodei had attended.

      This meeting occurred amidst a unique dispute in the relationship between Washington and the tech industry. In July 2025, Anthropic signed a $200 million contract with the Pentagon, becoming the first AI lab to have its models approved for use within the DOD's classified networks. However, as negotiations on deploying Claude on the department's GenAI.mil platform began in September, discussions collapsed. The Pentagon insisted that Anthropic provide unrestricted access to its models for all legal purposes. Anthropic firmly stated that its AI would not be utilized in fully autonomous weapon systems that target without human oversight and would not be used for domestic mass surveillance of Americans.

      In response, Defense Secretary Pete Hegseth labeled Anthropic a "supply chain risk to national security" in late February 2026, a classification typically reserved for companies linked to foreign threats. This designation, communicated to Anthropic’s leadership on March 5, required defense contractors to verify that they were not utilizing Anthropic's models in military work. Trump reinforced this action with his Truth Social directive. According to Anthropic, as argued in subsequent legal challenges, the designation was unprecedented; US District Judge Rita Lin criticized it in a ruling that granted Anthropic a preliminary injunction in late March, noting it seemed aimed not at a legitimate national security threat but at penalizing the company for "bringing public scrutiny to the government’s contracting stance," describing it as "classic illegal First Amendment retaliation."

      The legal situation is divided. A federal appeals court in Washington D.C. declined Anthropic’s request to temporarily block the supply chain risk designation on April 8. Meanwhile, Judge Lin’s preliminary injunction in a separate but related case prevents enforcement of Trump’s Truth Social ban on Claude across the wider government. Consequently, while Anthropic is barred from Pentagon contracts, it may continue collaborating with other government agencies as both cases progress. The DOD has still utilized Claude during the US-Iran war that began before the blacklist was enforced.

      The apparent shift in the White House’s stance may be attributed to Mythos. Portions of the intelligence community and the Cybersecurity and Infrastructure Security Agency (CISA) have been testing the model. The Office of Management and Budget at the White House is establishing protocols for federal agencies to access a controlled version of the model. Treasury Secretary Bessent's attendance at the meeting on Friday was interpreted by sources close to the negotiations as an indication that the economic and financial security arguments for access to Mythos have reached the top levels of the administration.

      As one source within the administration told Axios, "It would be grossly irresponsible for the US government to deny itself the technological advancements that this new model offers. It would be a gift to China."

      Whether the relationship between Anthropic and the Pentagon can resume remains unclear. Trump's comments on Tuesday alluded to discussions that appeared promising but did not culminate in a deal. The appeals court ruling regarding the supply chain risk designation is still in effect, and Hegseth has not retracted his stance. Anthropic has