HHS plans a sole-source Claude contract while preparing a wider AI competition

The Department of Health and Human Services (HHS) intends to award Anthropic a contract for Claude services without an open competition while it prepares a broader purchase of enterprise AI tools. The department set out that plan in a special notice posted on the federal procurement site SAM.gov on 19 October, FedScoop reported. The proposed award would keep an existing deployment available to department personnel rather than select a supplier through the forthcoming competition.

Claude access for HHS personnel

The interim contract would maintain Claude for Enterprise and Claude for Government for authorised HHS personnel until the department awards its planned multiple-award blanket purchase agreement, ExecutiveGov reported. A blanket purchase agreement provides a vehicle for subsequent purchases. HHS’s proposed agreement would involve multiple awards, in contrast with its proposed award to Anthropic alone for the existing service.

HHS made Claude available across its workforce in December 2025. Anthropic’s discounted arrangement under the General Services Administration’s OneGov initiative offered agencies access for $1 a year and is due to expire on 31 October. Those dates place the proposed continuity contract between a department-wide deployment and the end of the discounted arrangement, while HHS works on the wider procurement.

The services covered by the proposal extend beyond a staff chatbot. They include drafting, summarisation, search, research and reporting, alongside document processing, data analysis and coding assistance. The proposed scope also includes reasoning models, agentic functions and access through a secure application programming interface, which would allow software to use the service directly. ExecutiveGov reported that the contract would preserve access within HHS’s existing infrastructure.

AI Affairs previously reported that Anthropic opened Claude for Government to agencies without seat fees. The HHS proposal concerns continued access to services already in use, ahead of a separate process for buying enterprise AI.

Anthropic’s model in the sole-source case

HHS based its sole-source approach on the need to continue using Anthropic’s proprietary Claude model family, according to ExecutiveGov. That distinction matters to the proposed purchase: other suppliers might offer chat, drafting or data-analysis tools, but the department is seeking to maintain its existing Claude deployment. The same account says HHS intends the services to operate under its approved security and data-handling requirements, including encryption in transit and at rest, US data residency and audit logging. Provisions for protected health information would apply where eligible.

Jessica Tillipman, associate dean for Government Procurement Law Studies at George Washington University, questioned whether continuity justified foregoing competition, FedScoop reported. She regarded the capabilities in the notice as fairly general chatbot services and said the proposal appeared to acknowledge dependence on Claude. In her account, difficulty or cost in changing suppliers would point to a procurement problem rather than establish that no other supplier could provide those services.

That objection does not remove HHS’s immediate need to serve personnel who already use Claude. It puts the burden on the department’s narrower proposition: whether maintaining this deployment requires an award to its current provider while a competition is prepared. The planned agreement offers a route to competing awards, but it does not make the proposed interim contract competitive.

The HHS enterprise AI agreement

HHS has named its planned competition the HHS Enterprise Frontier SI Models and Advanced Features multiple-award BPA. Its intended use is to purchase enterprise AI services through awards to more than one supplier. For now, the notice places that competition after the proposed Anthropic contract. The department is therefore approaching the two purchases differently: continuity of a named model for current users, followed by a competitive vehicle for broader access.

Claude’s use has also varied within the department. The Food and Drug Administration began moving its Elsa chatbot from Claude to Google’s Gemini earlier in 2026 following what FedScoop described as a governmentwide ban on Anthropic. HHS’s proposal, by contrast, concerns continued Claude access for authorised personnel under an existing department deployment.

HHS set 21 October as the deadline for responses to the sole-source notice.