Pentagon reaches agreements with top AI companies, but not Anthropic
Sign up now: Get ST's newsletters delivered to your inbox
Anthropic, which has been in dispute with the Pentagon over guardrails for how the military could use its AI tools, was not among the seven companies listed.
PHOTO: REUTERS
WASHINGTON - The Pentagon said on May 1 it had reached agreements with seven leading artificial intelligence (AI) companies to deploy their advanced capabilities on the Defence Department’s classified networks.
The statement notably excludes Anthropic, which has been in dispute with the Pentagon over guardrails for how the military could use its AI tools.
The Pentagon labelled the AI start-up, which has its tools widely used across the Department of Defence, a supply chain risk earlier in 2026, barring the Pentagon and its contractors from using them.
SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft and Amazon Web Services will be integrated into its secret and top-secret network environments, providing more military access to their products for use on sensitive topics, the Pentagon said in a statement.
The lesser-known Reflection AI, which raised US$2 billion (S$2.54 billion) in October 2025, is backed by 1789 Capital, a venture capital firm in which US President Donald Trump’s son, Mr Donald Trump Jr, is a partner and investor.
Since the Pentagon deemed Anthropic’s products a “supply-chain risk” in March and the two sides became embroiled in a lawsuit, the military has expressed increasing interest in AI startups.
Faster process
Since the blow-up, newer AI entrants have said the military has sped up the process of incorporating them onto secret and top-secret data levels to less than three months. The process previously took 18 months or longer.
By expanding AI services offered to troops, who use it for planning, logistics, targeting and in other ways to streamline huge operations and perform more quickly, the Pentagon said in its statement it will avoid “vendor lock”, a likely nod to its overdependence on Anthropic or other dominant service providers.
Pentagon staffers, former officials and information technology (IT) contractors who work closely with the US military have told Reuters they were reluctant to give up Anthropic’s AI tools, which they view as superior to alternatives, despite orders to remove them over the next six months.
AI has become increasingly important for the US military. The Pentagon’s main AI platform, GenAI.mil, has been used by over 1.3 million Defence Department personnel, the agency noted in its release, after five months of operation.
Google, which is already used within the Pentagon, has signed a deal enabling the Department of Defence to use its AI models for classified work, a source told Reuters earlier this week.
Defence Department chief technology officer Emil Michael, on May 1, told news outlet CNBC that Anthropic is still a supply chain risk, but that Mythos, the company’s AI model with advanced cyber capabilities, is a “separate national security moment”.
While numerous companies and public and private entities have gained access to a Mythos preview product to help secure their IT infrastructure against future cyberattacks, it is unclear if the Pentagon is part of that programme.
US President Donald Trump said last week that Anthropic was “shaping up” in the eyes of his administration, opening the door for the AI company to reverse its blacklisting at the Pentagon. REUTERS
