Anthropic, the artificial intelligence company behind Claude AI, has announced a significant expansion of its presence across the US government. The company secured an agreement with the US General Services Administration (GSA), extending access to its AI products for federal, state, local, tribal, and territorial government organizations. This development unfolds even as Anthropic remains embroiled in a legal battle with the Pentagon, which has designated the company a national security supply-chain risk.
Details of the New GSA Agreement
Under the agreement, eligible government organizations will receive a substantial 40% discount on token-based usage of Claude Enterprise, Claude for Government, and the Claude API. Anthropic estimates this deal could save the US government up to $1 billion. Agencies will operate on a usage-based pricing model, eliminating minimum spending commitments. Access to Claude AI products will be available directly through Anthropic, via authorized resellers, and eventually through major cloud platforms.
Crucially, the agreement also includes provisions for authorized environments suitable for sensitive but unclassified work, broadening the scope of potential applications within government bodies. This new deal builds on an earlier offer that provided federal agencies access to Claude for Enterprise and Claude for Government for a nominal $1 per user, an offer set to expire on October 31, 2026.
Growing Government Adoption of Claude AI
Anthropic reports that approximately 500 US federal, state, and local government organizations currently utilize Claude AI. The company highlights a rapid increase in adoption, with weekly active users of Claude for Government growing eightfold since February 2026. Usage across federal civilian agencies has seen an even more dramatic twentyfold increase over the past year.
The company provided examples of Claude AI's impact, citing a four-week pilot program involving 34 developers across 10 Utah state agencies, where 71% of participants reported saving at least 20 hours. Another instance involved a state government team using Claude to respond to a ransomware attack in just eight minutes, a task that typically takes two to three days. Anthropic provided these figures and examples, which have not been independently verified.
The Ongoing Pentagon Dispute
Despite this expanding government footprint, Anthropic's relationship with the US administration is complicated by an unresolved dispute with the Pentagon concerning military AI safeguards. In February 2026, Anthropic refused to remove restrictions designed to prevent its technology from being used for mass domestic surveillance and fully autonomous weapons. Following this, the Trump administration directed federal agencies to cease using Anthropic's technology, and in March, the Pentagon formally designated the company a national security supply-chain risk.
The legal challenges have seen varied outcomes. In August, a California federal court struck down a related government action, citing unlawful retaliation. However, a US appeals court upheld the Pentagon’s designation on September 25, leading Anthropic to seek a rehearing on October 13. The new GSA agreement, while expanding access to Anthropic's AI, does not resolve this ongoing legal battle or overturn the Pentagon's designation. In a related development, US President Donald Trump hosted Anthropic CEO Dario Amodei for a dinner meeting in September amid the dispute.