After a dramatic turn of events following Anthropic’s feud with the US Department of Defense, the company’s CEO, Dario Amodei, has resumed negotiations with the Pentagon regarding the military’s access to use the company’s AI models. According to a report by the Financial Times, the Anthropic CEO is making efforts to resolve a high-profile standoff with the Trump administration that escalated last week, raising concerns about the AI safety guardrails governing the use of AI for national security purposes.
Anthropic, US Govt resolution concerns
Based on the report, it is said that discussions have restarted between both sides in a bid to reach a compromise agreement that would allow the Pentagon to resume using Anthropic’s AI technology while avoiding a formal blacklist designation. If a successful deal could be established, it could restore access for military applications under mutually acceptable safeguards, protecting Anthropic’s business interests and federal relationships.
Amodei has reportedly discussed the situation with key investors and partners, including Amazon CEO Andy Jassy, as well as venture firms like Lightspeed and Iconiq, emphasising the company’s commitment to finding a workable solution without compromising core principles. So far, Anthropic has positioned itself as the firm prioritising democratic values and robust safety measures at the expense of lucrative defence contracts.
However, neither Anthropic nor the Pentagon has officially commented on the matter.
US relying on Anthropic’s AI model for internal purposes
Earlier, talks between Amodei and Emil Michael, Under Secretary of Defense for Research and Engineering, had broken down after Anthropic had insisted on firm assurances in the contract that its technology, particularly Claude AI models, would not be used for mass domestic surveillance of US citizens or the deployment of fully autonomous lethal weapons.
The Pentagon, however, declined to provide such ironclad restrictions, prompting Defense Secretary Pete Hegseth to designate Anthropic a potential “supply chain risk” — a label typically applied to foreign entities. This led President Trump to order federal agencies to stop using Anthropic tools, effectively blacklisting the company from government contracts and pressuring defence contractors to cancel ties.
The fallout followed Anthropic’s refusal to drop its ‘ethical red lines’, contrasting sharply with rival OpenAI’s recent Pentagon deal following the spat, which included more flexible language around “lawful purposes” and recently drew criticism from Amodei as “safety theatre.” OpenAI CEO Sam Altman had to reword the contract after the company’s ChatGPT chatbot saw massive uninstallations.
What remains to be seen is how the Pentagon goes about accommodating the concerns raised by Anthropic in the deal and whether it will drop the new contract drawn up with rival OpenAI in the coming weeks.