OpenAI is set to significantly alter its pricing structure for U.S. government agencies, shifting from a nominal $1-a-year pilot program to a usage-based model with a substantial 50% discount on standard rates. This new agreement, which takes effect on October 1st and will span over two years, aims to facilitate broader adoption of AI technologies within the public sector.
Transitioning from a Landmark Pilot Program
Last year, OpenAI joined the General Services Administration’s (GSA) OneGov procurement platform, enabling federal agencies to acquire AI software in bulk. The initial $1-a-year deal was lauded as the most significant technology product discount ever negotiated by the U.S. government’s central procurement body. During this pilot phase, an impressive 3.5 million federal employees utilized ChatGPT, reportedly leading to cost savings of $1.4 billion.
While the new pricing represents an increase from the symbolic $1 annual fee, the 50% discount ensures that government entities can still explore and deploy AI tools, including OpenAI’s latest models like GPT-6 Astra, at a considerably reduced cost. Notably, this discount will soon extend beyond federal agencies to include state, local, and tribal governments.
Competitive Landscape and Legal Battles
This move follows a similar $1-a-year collaboration announced by competitor Anthropic PBC shortly after OpenAI’s new contract signing. Anthropic, the creator of the Claude chatbot, had also actively promoted its AI products to federal workers, with CEO Dario Amodei delivering a keynote speech in Washington D.C.
However, Anthropic’s government business quickly encountered significant hurdles. A dispute arose with the Department of Defense concerning Anthropic’s insistence on including security clauses related to surveillance and autonomous strike systems in national security contracts. This conflict led to U.S. Secretary of Defense Pete Hegseth designating the company as a supply chain risk.
The situation escalated when former President Donald Trump publicly called for a halt to Anthropic’s products across all government agencies, citing the company’s alleged “woke” and politically correct stance. This intervention prompted Anthropic to file a lawsuit against the U.S. government.
Ongoing Legal Proceedings
The legal ramifications of Anthropic’s dispute are still unfolding. Initially, the GSA removed Anthropic from its official AI models platform. However, a federal judge in California issued a temporary restraining order, leading the GSA to retract its decision.
More recently, the same California judge ruled that the Trump administration could not prohibit federal agencies from using Anthropic’s AI. The ruling clarified that while the Department of Defense could disable the technology for its own operations, it lacked the authority to mandate a universal ban across other government bodies. A separate lawsuit remains pending in Washington D.C., where a three-judge panel is yet to rule on another legal claim.
Meanwhile, OpenAI announced in February that it had reached an agreement to provide AI tools to the Department of Defense, indicating a different trajectory in its government partnerships compared to Anthropic’s recent challenges.
