Archive · 2 Sep 2026
Pentagon Launches AI Portal Without Anthropic’s Claude Model
The Pentagon created a secure website where military workers can use several powerful AI tools.
The tools include Gemini, ChatGPT and Grok.
More than half of the Defense Department’s roughly three million workers have joined the system.
Anthropic’s Claude was expected to be included, but it was not.
Anthropic wanted promises that its AI would not be used for mass surveillance of Americans or lethal autonomous weapons.
The Pentagon did not agree to those limits.
It later labeled Anthropic a national security supply-chain risk.
Anthropic sued, and a judge said the Pentagon’s actions were illegal and violated the company’s free-speech rights.
However, the ruling did not require the Pentagon to buy Anthropic’s technology.
The Pentagon launched GenAI.mil, a secure portal for commercial AI models on government systems.
The platform offers Google’s Gemini, OpenAI’s ChatGPT and xAI’s Grok to Defense Department personnel.
More than 1.7 million of roughly three million personnel have reportedly been onboarded.
Anthropic was excluded after seeking limits on mass surveillance and lethal autonomous weapons.
A federal judge ruled the Pentagon’s designation of Anthropic as a national security supply-chain risk was illegal and baseless.
- Who
- The U.S. Department of Defense, Google, OpenAI, xAI and Anthropic are central to the dispute.
- What
- The Pentagon launched GenAI.mil with Gemini, ChatGPT and Grok, but not Anthropic’s Claude.
- When
- Google’s products launched first; ChatGPT and Grok were added on August 31, and the portal expanded four days after the court ruling described in the article.
- Where
- The platform operates on U.S. government systems, while Anthropic’s lawsuit was heard in federal court in San Francisco.
- Why
- The Pentagon sought unrestricted access to commercial AI, while Anthropic sought limits on mass surveillance and lethal autonomous weapons.
Pentagon’s Position
Anthropic’s Position
Limits on military use
Pentagon’s Position
The Defense Department wanted commercial AI access without contractual restrictions on how the technology could be used operationally.
Anthropic’s Position
Anthropic sought guardrails barring mass surveillance of Americans and lethal autonomous weapons.
Vendor selection
Pentagon’s Position
The Pentagon maintains that procurement discretion allows it to choose other vendors and does not require it to purchase Anthropic’s technology.
Anthropic’s Position
Anthropic argued that the Pentagon’s treatment of the company was unlawful retaliation after it refused the requested terms.
AI safety commitments
Pentagon’s Position
Military officials have emphasized platform security certification, including Impact Level 5, while not publicly addressing specific uses such as targeting or decisions involving force.
Anthropic’s Position
Anthropic’s position reflects its stated concern that published safety commitments should apply even when the customer is the military.
- Platform
- GenAI.mil is a secure portal for commercial frontier AI models.
- Available models
- Google Gemini, OpenAI ChatGPT and xAI Grok are available.
- Potential users
- The platform is intended for roughly three million Defense Department personnel.
- Current adoption
- More than 1.7 million unique users have reportedly been onboarded.
- Security level
- The tools were described as certified to Impact Level 5.
- Excluded model
- Anthropic’s Claude was not added to the platform.
- Court ruling
- Judge Rita Lin ruled that the Pentagon’s supply-chain-risk designation and related actions against Anthropic were illegal and baseless.