Anthropic CEO Dario Amodei has reportedly opened last-minute talks with the US Department of Defense to secure continued access to Pentagon contracts as the company faces the prospect of a supply chain threat posed by the Trump administration.
Amodei has held talks with Emil Michael, the US undersecretary of defense for research and engineering, to finalize the terms of the military’s use of Anthropic artificial intelligence models, the Financial Times reported, citing people familiar with the matter.
According to the report, the new agreement will allow the Pentagon to continue to use the company’s technology and could prevent formal designations that would force contractors in the defense supply chain to cut ties with the AI developer.
The talks followed a dramatic breakdown in talks last week. Michael reportedly accused Amodei of “lying” with the “God complex” while the dispute ended after the two sides failed to agree on language Anthropic needed to prevent misuse of its technology.
related to: Former OpenAI Researcher’s Hedge Fund Reveals Bitcoin Miners’ Big Bets in New SEC Filing
Pentagon talks stall over general data analysis clause
In an internal memo to staff seen by the FT, Amodei wrote that near the end of the negotiations, the Pentagon offered to accept broader terms from Anthropic if the company removed the restrictive “data analysis” clause. He said the phrase was intended to protect against potential internal mass surveillance, an anthropic scenario as a red line, along with the use of AI in lethal weapons.
The dispute escalated after Defense Secretary Pete Hegseth warned that Anthropic could be designated a supply chain risk, a move that would effectively bar the company from US military procurement networks.
The opposition comes despite Anthropic’s existing ties to the defense sector. The company was awarded a contract worth up to $200 million by the US Department of Defense in July 2025, becoming the first AI provider to use its models in classified environments and by national security agencies.
As Cointelegraph reported, the US military even used Anthropic’s Claude AI model to support a major airstrike on Iran hours after President Donald Trump ordered federal agencies to stop using the company’s systems.
related to: Mining companies move deeper into AI, HPC as MARA can sell Bitcoin
Tech groups warn that the danger label could hurt US AI leadership
Meanwhile, in a letter to Trump on Wednesday, tech groups warned that labeling a domestic AI company a supply chain risk could undermine America’s leadership in AI. These groups argue that treating an American technology company “as a foreign competitor rather than an asset” could stifle innovation and weaken America’s ability to compete with China in the global AI race.
Signatories included the Software and Information Industry Association, TechNet, the Computer and Communications Industry Association, and the Software Business Alliance. The organizations represent hundreds of US tech companies, including AI chipmaker Nvidia, Google’s Alphabet and Apple.
Magazine: Bitcoin May Take 7 Years to Upgrade to Post-Quantum – BIP-360 co-author
