SI Agent News
ConfirmedClaudePolicySecurityBusiness

Anthropic expands government partnership with 40% discount on Claude

The new agreement with the GSA provides federal, state, and local agencies with significant cost savings to adopt AI for critical infrastructure and cybersecurity.

$1 billionestimated government savings

The short version

Why it matters

This move lowers the financial barrier for government agencies to adopt AI tools that handle sensitive, unclassified work. By offering a pay-as-you-go model without minimum spends, it allows smaller agencies to experiment with AI on tight budgets.

Expanding the reach of Claude

Anthropic is broadening its partnership with the General Services Administration (GSA) to make its AI models more accessible to public sector entities. This new agreement provides a 40% discount on several product surfaces, including Claude Enterprise, Claude for Government, and the Claude API. The discount applies to agencies buying directly from the company as well as those purchasing through cloud providers in the coming months.

The scope of this partnership is wide, extending beyond federal agencies to include state, local, tribal, and territorial governments. This expansion is designed to help these organizations manage important systems that often operate on limited budgets and face long procurement cycles. By reducing the cost of entry, Anthropic hopes to help these partners apply AI to difficult problems without requiring a large upfront financial commitment.

The growth in public sector adoption is already significant, with approximately 500 US government organizations currently using the technology. Anthropic reported that weekly active users for Claude for Government have increased about eightfold since February. Furthermore, usage across federal civilian agencies has seen a 20-fold increase over the past year, highlighting a growing trend toward AI integration in public service.

Practical applications in government

Government agencies are using Claude to address a variety of operational challenges, ranging from administrative tasks to complex technical maintenance. For example, some departments are using the tool to manage backlogs, organize records, and handle various forms. It is also being utilized to modernize old code, which is a common hurdle for agencies maintaining legacy systems.

Specific use cases highlight the efficiency gains for public servants. In Utah, 34 developers across 10 state agencies used Claude Code during a pilot program, where 80% of participants had a working environment ready within 30 minutes. By the end of a four-week period, 71% of those developers reported saving more than 20 hours of work.

Other agencies have used the technology for high-stakes technical tasks. One state CIO confirmed that their team responded to a ransomware attack in just 8 minutes, compared to the usual 2-3 days, thanks to Claude. Additionally, a federal security team used Claude Mythos to identify and fix a vulnerability in a critical production system that a human penetration test had missed the previous month.

Security and cloud infrastructure

To meet the rigorous requirements of the public sector, Anthropic offers Claude for Government, which provides near-feature parity with commercial versions at a FedRAMP High authorization level. This authorization allows agencies to use the tool for their most sensitive unclassified work. The product was made generally available on September 30 to support these specific security needs.

Agencies that prefer to build their own custom tools can access Claude through established cloud platforms like Amazon Bedrock or Google Cloud’s Vertex AI. On Amazon Bedrock, the models are authorized for FedRAMP High, and certain models are approved for DoD IL4 and IL5 workloads in AWS GovCloud. On Vertex AI, Claude is also authorized for FedRAMP High status.

Anthropic is continuing to pursue further certifications to expand these capabilities. Claude Enterprise is currently in the process of obtaining FedRAMP Moderate authorization through the FedRAMP 20x program. This initiative is part of a broader effort to bring the same innovative technology available to the private sector to government users while maintaining strict security protocols.

Implementation and next steps

To ensure that government agencies can successfully integrate the technology, Anthropic is launching a dedicated program to share best practices. These programs are modeled after the strategies used by their private sector customers to help agencies maximize their use of Claude. The company plans to begin sharing these best practices with public sector partners in the coming weeks.

The new agreement is structured to be flexible for different types of government departments. Agencies are not required to meet a minimum spend commitment, which allows them to pay only for the usage they actually consume. This pay-as-you-go model is intended to support the work of organizations that run some of the country's most important systems.

Interested departments can begin purchasing Claude at the OneGov discount through the GSA's Multiple Award Schedule immediately. Anthropic's Public Sector team is available to discuss specific options for departments that already use the tool or are considering it for the first time. This partnership marks a significant step in making advanced AI capabilities available to the broader public infrastructure.

What to watch

Try it

Check the GSA Multiple Award Schedule to see how the OneGov discount applies to your specific agency's procurement needs.

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by Anthropic · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news