HHS to make sole-source award to Anthropic ahead of broader AI competition

The Department of Health and Human Services said it’s planning to make an award to Anthropic for Claude services without an open competition as an interim plan for continuation of services.

In a notice Monday on SAM.gov, HHS disclosed its intent to sole-source an award to the artificial intelligence developer “for a limited continuity of services award” while the department prepares to make an award through a coming competitive process. 

In the meantime, the agency argued that Anthropic’s Claude is the only tool that can fulfill its needs — including chat, drafting, summarization, data-analysis workflows, application access, and administrative controls.

“The capability is specific to one provider: Claude is Anthropic’s proprietary model family, and the requirement is continuity of the existing Claude deployment already,” the notice said.

It’s difficult to say specifically which agreements the sole-source would continue based on the notice alone, but it is public knowledge that the agency uses Claude.

HHS rolled out Claude to its workforce in December 2025, following the company’s offer to provide its services to agencies at $1 a year and cited use cases for Claude in its AI inventory for 2025. 

The department also was using Claude in the Food and Drug Administration’s Elsa chatbot, but as reported by Fierce Biotech, the agency began transitioning that tool to Google’s Gemini earlier this year following the Trump administration’s governmentwide ban on Anthropic that was a response to the clash between the company and the Pentagon. 

The legal battle in that dispute is ongoing. But while the Pentagon appears to still be at odds with Anthropic, the company seems to be on decent terms with officials elsewhere in government.

Notably, the sole source notice also comes as Anthropic’s discounted deal to agencies through the General Services Administration’s OneGov initiative is set to come to an end Oct. 31. Like other providers, that end date is an extension of its original deal. And following the conclusion of that deal, at least OpenAI has announced a new plan for consumption-based pricing.

HHS and Anthropic didn’t provide responses to FedScoop’s inquiries.

The notice is also raising eyebrows.

Jessica Tillipman, associate dean for Government Procurement Law Studies at George Washington University, told FedScoop the notice appears to be an admission that the agency is dependent on Claude.

If it’s too much of a pain or too cost prohibitive to switch to a new vendor, Tillipman said the agency negotiated poorly despite all the warnings. Tillipman has previously written about the risk of dependency, or so-called vendor lock-in, with respect to the budding area of AI procurement. 

Tillipman also noted that the description in HHS’s notice seems to outline fairly general chatbot services, and that the department’s assessment that only Claude could provide those services is a questionable use of sole-sourcing.

“To me, it just reeks of … ‘this is going to be a pain, so let’s just keep it going,’ but that’s not a normal justification,” Tillipman said.

In the notice, the department said that it eventually plans to purchase enterprise AI services through a competitive multiple-award, blanket purchase agreement, calling that competition the “HHS Enterprise Frontier SI Models and Advanced Features multiple-award BPA.” It wasn’t immediately clear whether the competition was active.