The president's comments follow a meeting with Anthropic representatives and a months-long standoff over military AI use
NurPhoto / Getty Images
President Donald Trump said it is possible the U.S. government may strike a deal to use Anthropic's AI models in the Department of Defense. The move would signal a potential thaw in the high-profile dispute between the administration and the AI company.
Speaking on CNBC's "Squawk Box," Trump said recent White House meetings had gone well and that Anthropic representatives were "shaping up" and could "be of great use." He also said that while he seeks "the smartest people," OpenAI has stepped in to fill the role Anthropic previously held.
Friday's White House gathering included White House Chief of Staff Susie Wiles, U.S. Treasury Secretary Scott Bessent, and Anthropic CEO Dario Amodei, who brought the company's newly unveiled Mythos model to the table. A White House spokesperson called the session "productive and constructive," per CNBC.
The standoff centered on Anthropic's refusal to give the Pentagon open-ended authority over its models, especially for autonomous weapons or surveillance of American citizens. The DOD responded by classifying Anthropic as a supply-chain threat, obligating military contractors to drop Claude from any government-facing work. Trump separately took to Truth Social to demand that all federal agencies cut ties with the company immediately.
The company sued the administration in California and Washington, seeking to undo the blacklist. A California judge issued a temporary halt to the ban, but the D.C. Circuit Court of Appeals has allowed the supply-chain designation to remain as the case proceeds, according to Yahoo Finance.
The legal fight has not stopped the Pentagon from using Anthropic's technology. The DOD continued to use Claude during operations in Iran, according to CNBC.
A $200 million Pentagon contract had been in place since July, but by September, efforts to run Claude on the DOD's GenAI.mil platform broke down without resolution.
Mythos, Anthropic's most capable model yet, was made available only to a curated set of partners given the sensitivity of its cybersecurity features. Its launch appears to have opened the door to renewed government engagement. Amodei had already been building that rapport in early April, when he joined other leading tech executives in a call with Bessent and Vice President JD Vance focused on AI and cyber preparedness.
Join 500,000+ readers who start their day with Quartz.
By subscribing, you agree to our Terms of Service and Privacy Policy.
