Anthropic moved back into a Washington fight on Tuesday after President Donald Trump said a deal for the company’s AI models inside the Department of Defense could still happen.
Speaking on CNBC’s “Squawk Box,” Trump said, “it’s possible” there will be an agreement that allows Anthropic technology to be used by the military. He also said:
“They came to the White House a few days ago, and we had some very good talks with them, and I think they’re shaping up. They’re very smart, and I think they can be of great use.”
The remarks marked a change in tone after months of conflict between Anthropic, the Pentagon, and the Trump administration. In March, the DOD labeled Anthropic a supply chain risk, saying the company’s technology could threaten U.S. national security.
The label forced defense contractors to certify that they were not using Anthropic’s Claude models in military work. Trump then told federal agencies to “IMMEDIATELY CEASE all use of Anthropic’s technology,” while adding that his administration would “not do business with them again.”
Pentagon keeps using Claude while Anthropic fights the blacklist and reopens talks
That hard line did not fully hold. The Pentagon kept using Claude during the war with Iran. Anthropic later sued the Trump administration in San Francisco and Washington, D.C., to reverse the blacklist.
Trump’s Truth Social directive has also been temporarily blocked by a federal judge. Talks between both sides then started opening again. Anthropic chief executive Dario Amodei met senior administration officials on Friday to discuss Mythos, the company’s new AI model with cybersecurity capabilities. White House chief of staff Susie Wiles and Treasury Secretary Scott Bessent attended that meeting.
A White House spokesperson allegedly called the discussion “productive and constructive.” Earlier this month, Anthropic announced Mythos and limited its release to a small group of companies because of the model’s cyber power. The company said it has been holding “ongoing discussions” with U.S. government officials about Mythos.
Mythos arrived after the worst point of Anthropic’s dispute with the DOD. The launch appears to have reopened the door to better ties with the administration.
Dario also joined an early April call with Bessent and Vice President JD Vance to discuss AI cyber readiness alongside other major tech CEOs. Anthropic signed a $200 million Pentagon contract in July, but negotiations over deploying Claude on the DOD’s GenAI.mil platform collapsed in September.
Banks rush toward Mythos as Anthropic prepares a wider European rollout
The company is facing pressure outside Washington too. In New York and Paris on April 21, Reuters reported that Anthropic plans to give European banks access to Mythos soon, citing three people familiar with the matter.
This comes as banks scramble to test the model after large U.S. banks received the first access. Cybersecurity experts see Mythos as a challenge for banks and their old technology systems, and those fears drove warnings from regulators and policymakers at last week’s International Monetary Fund spring meeting in Washington.
Scott Keipper, EY’s Americas Financial Services Technology Consulting Leader, said the speed of the technology is outrunning the governance, operating models, and control systems most banks were built to handle.
He said that the gap is widening the distance between finding risk and fixing it. Keipper also said banks need to move past one-time cybersecurity fixes and instead build AI into risk management across technology, operations, governance, and oversight.
One person familiar with the matter told Reuters that Anthropic wants to expand Mythos access to European and UK banks, along with other organizations. That person said security checks are part of the rollout process.
Another person allegedly said European banks could get access within days, while the first person said the timeline could still take days or weeks. Bloomberg had already reported that Anthropic was preparing to release Mythos to UK financial institutions soon.
