Pentagon Pursues New AI Models as Anthropic Contract Falls Apart
Key Points
- Pentagon is engineering multiple LLMs for government‑owned environments.
- Anthropic's $200 million DoD contract ended over access and usage disputes.
- Anthropic wanted prohibitions on mass surveillance and autonomous weapons.
- OpenAI and xAI (Grok) have secured separate agreements with the Pentagon.
- Defense Secretary labeled Anthropic a supply‑chain risk, restricting collaborations.
- Anthropic is contesting the supply‑chain risk designation in court.
After a contentious split, the Pentagon is developing its own large‑language‑model tools to replace Anthropic's AI. The Department of Defense announced engineering work on multiple LLMs for government‑owned environments and expects operational use soon. Anthropic's $200 million contract collapsed over disputes about unrestricted access, mass‑surveillance prohibitions, and autonomous weapon use. While OpenAI and Elon Musk’s xAI have secured separate agreements with the Pentagon, Defense Secretary Pete Hegseth labeled Anthropic a supply‑chain risk, a restriction that Anthropic is now challenging in court.
Pentagon's Shift Away from Anthropic
In a recent briefing, Cameron Stanley, the Pentagon's chief digital and AI officer, confirmed that the Department of Defense is actively pursuing multiple large‑language‑model (LLM) systems for placement in government‑owned environments. Engineering work on these models has already begun, and officials expect the new tools to be available for operational use in the near future.
The move follows a dramatic breakdown in the relationship between Anthropic and the Pentagon. Anthropic's $200 million contract with the Department of Defense fell apart after the two parties could not agree on the extent of the military's unrestricted access to the company's AI technology. Anthropic sought to embed a contractual clause that would prohibit the Pentagon from using its AI for mass surveillance of Americans or for deploying weapons that could fire without human intervention. The Pentagon did not accept these conditions, leading to the contract's termination.
With Anthropic out of the picture, the Pentagon has turned to other AI providers. OpenAI has entered into its own agreement with the Department of Defense, and the agency has also signed a deal with Elon Musk’s xAI to integrate the Grok model into classified systems. These new partnerships underscore the Pentagon's intent to maintain advanced AI capabilities while avoiding reliance on Anthropic.
Defense Secretary Pete Hegseth has designated Anthropic as a supply‑chain risk, a status typically reserved for foreign adversaries. This designation bars companies that work with the Pentagon from collaborating with Anthropic. In response, Anthropic is challenging the supply‑chain risk label in court, seeking to overturn the restriction.
The Pentagon's strategy reflects a broader effort to ensure that its AI tools are housed within secure, government‑controlled environments and that the agency retains full operational authority over the technology. By developing its own LLMs and securing agreements with other leading AI firms, the Department of Defense aims to sidestep the legal and policy hurdles that derailed its relationship with Anthropic.