Pentagon says it has stopped using Anthropic products
A Pentagon official said the department had ceased using Anthropic products, after Claude was still in use as recently as last week. The AI tool had supported intelligence work and military operations against Iran.

Key takeaways · 4
- 01
The Pentagon official’s statement says Anthropic products are no longer in use by the department.
- 02
Claude was still being used last week for research, analysis, intelligence gathering and military operations against Iran.
- 03
Claude was integrated into Maven Smart System, and that integration would make it difficult to remove quickly.
- 04
Anthropic says it will provide models and engineer support during the transition, for as long as necessary and permitted.
A shift from planned exit to reported cessation
A department official said Monday that the Pentagon had ceased using Anthropic products.[1] The statement follows a February decision by Defense Secretary Pete Hegseth to designate Anthropic a supply-chain risk.[1] The Pentagon had said it would stop using Anthropic by late August.[1] The available statements therefore include both a previously stated end-of-August timetable and a later official account that use had ceased; they do not explain how the timetable changed or give a specific date for the cessation.[1]
Claude had operational and analytical roles
Claude was still in use by the department as recently as last week.[1] Its uses included research, analysis and intelligence gathering, as well as military operations against Iran.[1] It was viewed as a critical tool for sorting through large quantities of information.[1] Anthropic’s products had been used by the US government and military since 2024, and the company was the first advanced AI firm to have tools deployed in government agencies doing classified work.[1] The evidence does not specify how many users or missions relied on Claude.
A dispute over guardrails and military access
The Pentagon had pressed Anthropic to remove Claude’s safety guardrails and give the military unfettered access to its AI products.[1] Anthropic refused, citing concerns about mass surveillance and autonomous weapons.[1] The company says it supports lawful national-security uses of AI, subject to two narrow exceptions, and believes those exceptions had not affected a government mission to that point.[2] Anthropic called the designation unprecedented and unlawful and sued the Trump administration to overturn it.[1] The evidence describes the disagreement but does not report an outcome in court.
Transition could involve embedded systems
Claude was embedded in Maven Smart System, a broader platform operated by Palantir that the Pentagon uses to organize intelligence and other data.[1] Its deep integration would have made Claude difficult to remove quickly; Kahn said removing integrated tools can be painful.[1] Anthropic says it will provide its models to the Department of War and national-security community at nominal cost, with continuing engineer support, during the transition for as long as necessary and permitted.[3] The company also says the designation does not affect access through its API, claude.ai or its products.[2]
For teams evaluating AI in sensitive workflows, this case highlights that product access, government contracting and integration into larger systems can be affected differently. It also shows why transition plans should account for both contractual restrictions and the practical work of replacing embedded tools.
Why it matters
Put this to work — one session a day, built for your industry.
Create a free account for a daily session — eight questions and one real-work challenge, on the news that affects your role.
Start freeHow this developed
5 October 2026
Pentagon says it has stopped using Anthropic products