The Pentagon said on Friday it had reached agreements with seven artificial intelligence companies to deploy advanced capabilities across its classified networks, as part of efforts to expand the range of AI providers working with the military.
In a statement, the department said firms including SpaceX, OpenAI, Google, NVIDIA, Reflection AI, Microsoft and Amazon Web Services would be integrated into its secret and top-secret network environments.
The move will enable broader military access to AI tools for use in sensitive operations, including planning, logistics and targeting, the statement said.
However, the announcement excluded Anthropic, which has been at odds with the Pentagon over safeguards governing the military’s use of its technology. The department earlier labelled the company a “supply-chain risk”, barring its use by the Pentagon and its contractors.
Despite the restriction, officials and contractors familiar with the matter told Reuters that Anthropic’s tools remain widely used and are viewed by some as more advanced than alternatives, complicating efforts to phase them out over the next six months.
The dispute has coincided with a broader push by the Pentagon to diversify its AI partnerships and avoid overreliance on a single provider. Officials said the onboarding process for new AI firms has been significantly accelerated, with integration timelines reduced from over 18 months to less than three.
The department said expanding AI services would help prevent “vendor lock” and enhance operational efficiency across large-scale military activities.
AI has become an increasingly critical component of US defence strategy. The Pentagon noted that its main AI platform, GenAI.mil, has been used by more than 1.3 million Defence Department personnel within five months of its launch.
Separately, a source told Reuters that Google has signed an agreement allowing the Pentagon to use its AI models for classified work.
Defence Department Chief Technology Officer Emil Michael said Anthropic continued to pose a supply-chain risk, although he described its advanced cyber-focused model, Mythos, as a “separate national security moment” due to its potential implications.
Meanwhile, Donald Trump said last week that Anthropic was “shaping up”, suggesting the possibility of a review of its current status with the department.
