The Pentagon Signs Deals with 8 AI Suppliers in One Fell Swoop, with Anthropic the Sole One Missed

According to OSINT Beating monitoring, the Pentagon has announced agreements with 8 technology companies to allow their AI tools and infrastructure to be deployed on the U.S. military's Secret Level (IL6) and Top Secret Level (IL7) classified network. The new signatories include Nvidia, Microsoft, AWS, and the startup Reflection AI, in addition to the previously contracted xAI, OpenAI, Google, and the added Oracle on the same day.

Key terms of the agreements: all companies agree that the Pentagon can use their technology for "any lawful purpose." Anthropic is not on the list. Previously, Claude was the only large model used by the Pentagon in classified environments, but CEO Dario Amodei refused to accept this standard, demanding written assurance from the government that it would not be used for large-scale domestic surveillance and lethal autonomous weapons. At the end of February, the Pentagon listed Anthropic as a supply chain risk, and Trump subsequently banned its participation in all government contracts, leading to a legal battle in federal courts between the two parties. Defense Secretary Hegseth referred to Amodei as an "ideological lunatic" in Congress.

However, The New York Times, citing insiders, reported that at least some of the new signatory companies, while accepting the "any lawful purpose" standard, privately obtained similar safeguards to those requested by Anthropic, involving restrictions on autonomous drones and domestic surveillance.

Pentagon CTO Emil Michael described the shift to open-source models as a new direction. Nvidia and Reflection AI provide open-source models, while Microsoft and AWS offer infrastructure, and xAI, OpenAI, and Google provide closed-source models. Michael stated that Chinese open-source models have already infiltrated many companies, necessitating American alternatives.

Despite the Pentagon's urgency to replace Anthropic with OpenAI and Google, the migration process has encountered technical issues, with intelligence analysts still using Anthropic's old models to perform daily tasks. Although Trump ordered Anthropic to be decommissioned by mid-year, Mythos has excelled in network attack and defense capabilities, leading The New York Times to report that this ban is almost certain to be revoked.