GSA offers 40% discount on Anthropic Claude for government agencies

Expansion of the OneGov deal

The General Services Administration (GSA) announced an expansion of its OneGov agreement with Anthropic on October 27. This new deal provides federal, state, local, tribal, and territorial government entities with a 40% discount on eligible usage of Claude artificial intelligence models. The move is designed to make advanced AI capabilities more accessible to a broader range of public sector organizations.

This new arrangement succeeds a previous OneGov promotion that gave agencies access to Claude for $1 per user, which was set to end on October 31. By expanding the deal, the GSA aims to provide a more sustainable and scalable way for government agencies to integrate AI into their daily operations. The new terms are available to all three branches of the federal government, including the executive, legislative, and judicial branches.

Anthropic estimates that this specific expanded agreement could save the government a total of $1 billion. The company noted that it is seeing significant growth in adoption across the public sector. Currently, approximately 500 federal, state, and local government organizations are already using Claude tools in their workflows.

Moving to consumption-based pricing

A key feature of the new agreement is the shift toward a consumption-based pricing model. Under these terms, government agencies will pay only for the AI capabilities they actually use rather than paying a flat fee for a set number of users. This approach ensures that costs are tied directly to the volume of work performed by the AI systems.

The GSA stated that this model is intended to scale with actual use, allowing for more flexibility in how different departments deploy the technology. There are no minimum purchase requirements under the new terms, which allows smaller agencies or specific projects to utilize the models without committing to large upfront costs. This mirrors a strategy the GSA has used with other providers, including OpenAI.

The OneGov initiative as a whole uses the government's collective purchasing power to negotiate standardized prices for commercial technology. By consolidating demand, the GSA can secure better rates for the federal government and taxpayers. This strategy has already generated more than $1.8 billion in savings since its inception.

Security and deployment options

Security remains a primary focus for government AI adoption, and the expanded agreement includes models in FedRAMP-authorized environments. These environments are designed for sensitive unclassified work, meeting the security standards required for many government tasks. Anthropic recently made Claude for Government generally available on September 30 with FedRAMP High authorization.

Agencies have multiple pathways to access these secure environments. They can purchase Claude through GSA’s Multiple Award Schedule, or they can access it through cloud platforms like Amazon Web Services (AWS) Bedrock and Google Cloud’s Vertex AI. Both of these cloud platforms offer FedRAMP High-authorized environments for government users.

For specific high-security needs, certain Claude models on Bedrock are approved for Department of Defense Impact Level 4 and 5 workloads in AWS GovCloud. This allows for a range of applications, from basic administrative tasks to complex software development. Anthropic has also announced a dedicated program launching in the coming weeks to help agencies implement these tools.

Growth in government AI usage

The data shows a rapid increase in how government workers are interacting with Claude. Anthropic reported that the number of weekly active users of Claude for Government has grown by about eight times since February. This suggests that the tool is becoming a staple for public sector employees across various departments.

The growth is even more pronounced within federal civilian agencies. Anthropic reported that usage across these specific agencies has grown about 20-fold over the past year. This trend highlights a significant shift in how the public sector is approaching the integration of generative AI into its infrastructure.

The expanded agreement covers several of Anthropic's core offerings, including Claude Enterprise and the Claude API. By providing access to these tools, the GSA is enabling agencies to automate a variety of tasks. These range from routine administrative work to more technical requirements like software development and data processing.

What comes next for OneGov

The GSA plans to continue its OneGov strategy by renewing agreements as existing offers expire. The goal is to maintain consistent access to modern technology while securing favorable governmentwide pricing. The agency did not specify a specific expiration date for the new Anthropic agreement, suggesting a long-term commitment to the partnership.

Agencies interested in the new pricing can find the necessary ordering details through the GSA’s OneGov portal. While direct purchasing is currently available for many entities, Anthropic expects cloud-provider purchasing options to become available in the coming months. This will provide even more flexibility for agencies that prefer to manage their AI through existing cloud contracts.

The GSA continues to work on making advanced AI capabilities, which the federal government has dubbed super intelligence, available across the board. By negotiating with major providers, the GSA aims to ensure that the government can keep pace with technological advancements. This ongoing effort is a core part of the broader initiative to modernize government services through commercial technology.