OpenAI Ends $1 Deal for U.S. Federal Agencies, Shifts to Half-Price Pricing
nashnova research
OpenAI will end its $1-a-year pilot with U.S. federal agencies on October 1, replacing it with a usage-based contract at 50% off standard pricing — a commercial milestone that lands just as rival Anthropic is locked in federal contract disputes and litigation.
What was the $1 deal about?
In August 2025, OpenAI joined OneGov — the federal AI procurement platform run by the General Services Administration (GSA) — at a symbolic $1 per year.
It was the largest discount OneGov had ever negotiated. During the pilot, roughly 3.5 million federal employees used ChatGPT, generating an estimated $1.4 billion in cost savings, according to GSA.
This means → the $1 was not charity. It was a land grab — build a 3.5-million-user habit first, then monetize.
How does the new deal work, and who is covered?
The new contract takes effect October 1 and runs over two years. Federal employees pay half the standard rate, billed by usage.
The discount extends beyond federal agencies to state, local, and tribal governments.
It covers OpenAI's newly released GPT-6 Astra model.
In plain terms = "free" becomes "half price," and the scope widens from federal to every level of U.S. government — OpenAI starts earning real revenue from the public sector.
Why has Anthropic fallen behind?
Days after OpenAI secured its GSA deal in 2025, Anthropic matched the $1 terms. CEO Dario Amodei personally flew to Washington to pitch Claude to federal workers.
Then Anthropic clashed with the Pentagon over safety provisions tied to surveillance and autonomous strikes. Defense Secretary Pete Hegseth flagged the company as a supply-chain risk; President Donald Trump posted on Truth Social demanding all federal agencies stop using Anthropic products.
Anthropic sued the U.S. government. A California court issued a preliminary injunction ordering GSA to reverse the de-listing, but a separate Washington case remains unresolved — the outcome is still uncertain.
This means → Anthropic's federal market access is stuck in court, while OpenAI used the same window to convert its pilot into a paying contract.
What does the divergence tell us?
OpenAI reached a deal with the Pentagon in February to supply AI tools; Anthropic was shut out over the same safety-clause disagreements.
This reflects a deeper split: AI companies willing to align with government security demands get contracts; those that hold their own safety red lines get excluded.
In plain terms = in the federal AI procurement game, OpenAI chose to sit at the table and Anthropic chose to go to court first — the short-term results speak for themselves.
市场有风险，内容仅供研究参考，不构成投资建议。