The Claude logo. Image: Anthropic
The US Department of Health and Human Services wants to keep its staff on Anthropic’s Claude without opening the work to rival bidders. In a notice of intent to award a sole source contract posted on SAM.gov on October 19, HHS says “the capability is specific to one provider” and that it needs “continuity of the existing Claude deployment already integrated into HHS infrastructure”.
The award would be a stopgap. HHS says it will cover the department only until it awards a competitive contract it calls the “HHS Enterprise Frontier SI Models and Advanced Features multiple-award BPA”, a blanket purchase agreement open to several suppliers. The name uses “SI”, the label Trump’s September 29 executive order told agencies to use in place of “AI”.
What HHS wants to keep
The notice, filed by the department’s IT buying centre under solicitation number 7571TE27R00001, asks for continued access to Claude for Enterprise and Claude for Government for “authorized HHS workforce” users. The list of covered features is long:
- Everyday work: general chat, reasoning, drafting and summarising, search, research, and file, document and data-analysis workflows.
- The top models: Anthropic’s “highest-capability/premium reasoning models and long-context modes”.
- Tools for builders: coding assistance, developer tools, agentic and “delegated-work” features, connectors, projects, memory, and API access.
- Controls: single sign-on, user and role management, feature controls, reporting and audit exports.
It also sets security terms: HHS data kept out of model training by default, encrypted storage and transfer, data kept in the continental US, a HIPAA business associate agreement for tools that handle patient health information, notice of any security incident within an hour, and secure deletion when the contract ends. The notice gives no value or length for the award. Other suppliers had until 3pm Eastern on October 21 to respond.
Why HHS says only Claude will do
HHS cites the federal rule that lets a buyer go to a single supplier when only one is reasonably available. Its case rests less on what Claude can do than on the fact that the department already runs it: Claude is “Anthropic’s proprietary model family”, the notice says, and HHS wants to keep the setup it has. HHS rolled Claude out to its workforce in December 2025.
Jessica Tillipman, associate dean for government procurement law studies at George Washington University, told FedScoop the notice reads like an admission that HHS depends on Claude, and that the services it describes are fairly general chatbot tools that other companies also sell.
“That’s not a normal justification”
Tillipman was blunt about the reasoning:
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.
Jessica Tillipman, George Washington University, to FedScoop
She added that if switching suppliers is too painful or too costly, the department negotiated poorly despite warnings about getting locked in to one AI vendor. Neither HHS nor Anthropic responded to FedScoop’s questions.
A different story at the Pentagon
The notice lands while another part of the government is trying to cut Anthropic off. In February, President Trump ordered federal agencies to stop using Anthropic’s technology, and Defense Secretary Pete Hegseth moved to brand the company a supply chain risk after it refused to drop limits on military use. An appeals court let the Pentagon keep Anthropic on its blacklist in September, the Pentagon said this month it had stopped using Claude, and Anthropic has asked the full court to rethink that ruling.
HHS has not been uniformly loyal either. FedScoop notes that the Food and Drug Administration, part of HHS, began moving its Elsa chatbot from Claude to Google’s Gemini earlier this year, and that Anthropic’s discounted deal for agencies through the General Services Administration’s OneGov programme ends on October 31.
Why it matters
One of the government’s biggest civilian departments is using a no-competition rule to say no other AI will do, at the same time as the Pentagon fights in court to keep Anthropic out. It is also an early test of how hard it is for agencies to switch AI suppliers once their staff rely on one, as the cheap introductory government deals come to an end.
Sources: HHS notice of intent to sole source to Anthropic, SAM.gov; FedScoop (Jessica Tillipman interview).
Latest Anthropic news
