Anthropic is cutting what the US government pays for Claude by 40%, under an expanded agreement with the General Services Administration announced in a blog post on Tuesday, October 27. For the first time, the discount extends beyond federal agencies to state, local, tribal and territorial governments as well.
The cut applies to Claude Enterprise, Claude for Government and the Claude API, covering token-based usage across every Claude model the government has access to, according to GSA's own release. There's no minimum order. Agencies pay only for what they actually use. Anthropic says the arrangement will save the government roughly $1 billion.
It replaces Anthropic's first OneGov agreement from August 2025, which had let federal agencies run Claude for a nominal $1. That deal expires October 31, Christian Robles reported for Nextgov/FCW. The new one runs 27 months, through January 31, 2029, per a GSA website cited by Nextgov.
A parallel deal with OpenAI
Anthropic isn't the only AI company GSA has struck a bargain with. OpenAI has a similar arrangement offering a 50% discount on token-based usage, also running 27 months, that began at the start of October, Nextgov reported. Last week, the Department of Health and Human Services said it planned a sole-source contract for Claude models specifically, Nextgov added, a sign that individual agencies are already picking sides between the two.
About 500 US federal, state and local government organizations now use Claude, Anthropic said. Weekly active users of Claude for Government have grown roughly eightfold since February, and usage across federal civilian agencies is up about 20-fold over the past year.
Some of the use cases Anthropic is pointing to are specific. Utah gave Claude Code to 34 developers spread across 10 state agencies. In a four-week pilot, 71% of them said it saved them 20 hours or more. Separately, a federal security team used Claude Mythos, a model Anthropic had held back on safety grounds, to find and patch a vulnerability that a human penetration test had missed. More than half of US states have taken part in Anthropic's cyber defense program, the company said.
Timed to an IPO filing
GSA says its broader OneGov program, which negotiates bulk software deals across agencies, has saved more than $1.8 billion with about $1.6 billion of that coming from what it calls "super intelligence agreements," a category that now presumably includes both the Anthropic and OpenAI contracts.
The discount news lands in the same week Anthropic filed to go public on Nasdaq. In materials for that filing, the company has told prospective investors that US government actions could affect its business, a risk disclosure that reads differently now that Washington is also one of its biggest customers.
For agencies, the clock that matters is October 31, when the old $1 pilot lapses and the 40% discount takes over. For Anthropic, the clock that matters is the IPO itself, where investors will be weighing how much of its growth is coming from deals like this one.
TBFiled by The Tech Business Desk
Big tech, startups, earnings, regulation and policy — the decisions that shape the products.
More from this desk →
Be the first to comment
Join the argument. No password, just your email or a passkey.