Anthropic was the first AI company to put its models on the classified cloud.

Judged as of 2026-02-28. True: dated announcements or reports show that Anthropic's models were available on a classified US government cloud or network (Secret or Top Secret level, for example through AWS or Palantir) before any other frontier AI model developer's models, such as OpenAI's, Google's or Meta's. Mostly true: Anthropic was first, or effectively tied for first, among frontier AI labs, but a minor detail is off, such as the exact network or the partner involved. Mixed: priority depends on the definition: Anthropic was first at one classification level or on one cloud, but another lab's models reached a classified environment earlier by a different route. Mostly false: Anthropic's models were on classified clouds early, but dated records show at least one other frontier AI company's models were deployed there before them. False: Anthropic's models had not been deployed on a classified cloud by that date, or several other AI companies clearly preceded them.

Set before the research began.

1. E1 Federal Government and Anthropic: Considerations for AI Innovation and Competition (IF13217) Congressional Research Service · 5 May 2026

   “In June 2024, Anthropic stated it was the first AI company to deploy frontier models in classified USG networks.”

2. E2 Anthropic teams up with Palantir and AWS to sell AI to defense customers TechCrunch · 7 November 2024

   “Claude became available on Palantir’s platform earlier this month and can now be used in Palantir’s defense-accredited environment, Palantir Impact Level 6 (IL6).”

3. E3 Anthropic teams up with Palantir and AWS to sell AI to defense customers TechCrunch · 7 November 2024

   “Meta recently revealed that it is making its Llama models available to defense partners, while OpenAI is seeking to establish a closer relationship with the U.S. Defense Department.”

4. E4 Amazon Bedrock launches with Claude 3.5 Sonnet in the AWS Top Secret cloud Amazon Web Services (AWS Public Sector Blog) · 13 January 2025

   “Amazon Bedrock is now generally available in the Amazon Web Services (AWS) Top Secret cloud.”

5. E5 Air-Gapped LLM Deployment: Running Frontier Models Offline GPUSmith · 21 July 2026

   “In May 2024, Microsoft announced it had deployed the GPT-4 large language model into an isolated, air-gapped Azure Government Top Secret cloud”

6. E6 Anthropic’s safety-first AI collides with the Pentagon as Claude expands into autonomous agents Scientific American · 21 February 2026

   “In late 2024 Anthropic made Claude available on a Palantir platform with a cloud security level up to “secret”—making Claude, by public accounts, the first large language model operating inside classified systems.”

7. E7 Announcing GPT-5.2 Availability in Azure for U.S. Government Secret and Top Secret Clouds | The Microsoft Cloud Blog Microsoft Cloud Blog · 15 January 2026

   “Today, we are excited to announce that GPT-5.2, Azure OpenAI’s newest frontier reasoning model, is available in Microsoft Azure for U.S. Government Secret and Top Secret cloud environments.”

A later secondary report (E5) dates a Microsoft GPT-4 deployment in a Top Secret cloud to May 2024. Anthropic stated in June 2024 that it was already deployed in classified networks (E1), but when that deployment happened is not in the evidence. So the evidence does not establish which company's models reached a classified cloud first.