On May 1, 2026, the Pentagon announced a major decision. It has allowed eight powerful artificial intelligence companies to work on its most secret computer networks. The eight companies are Amazon Web Services, Google, Microsoft, NVIDIA, OpenAI, SpaceX, Oracle, and Reflection AI. However, one well-known AI company was not included: Anthropic, the San Francisco company that created Claude.
This exclusion is not a mistake. It is part of an unusual conflict between Anthropic and the U.S. government. In February, the Trump administration called Anthropic a "supply chain risk" because the company refused to let Claude be used for fully autonomous weapons or to spy on Americans. Anthropic went to court and won an injunction against this decision in March. However, the Pentagon did not include Anthropic in its new agreement.
The eight approved companies can now use their AI models on classified Department of Defense networks. Defense Secretary Pete Hegseth calls this plan an "AI-first fighting force." Pentagon CTO Emil Michael explained that using multiple companies prevents the military from depending too much on one partner. The AI models will work through GenAI.mil, the Pentagon's generative AI platform. More than 1.3 million Department of Defense workers already use this service. The expansion into classified networks represents a historic moment, as commercial AI models have never before been allowed in such sensitive military systems.