The GSA OneGov Program and USAi Platform are two U.S. General Services Administration (GSA) initiatives announced over August 6–21, 2025 to broaden federal-agency access to generative AI. USAi.gov is a shared-services evaluation platform, and the OneGov deals provided several AI vendors' enterprise products to agencies at a nominal price of $1 per agency per year. The source is a GSA report dated August 14, 2025.
USAi platform
USAi.gov, announced on August 14, 2025, is a generative AI evaluation suite offered free to federal agencies. Its features include chat-based AI, code generation, and document summarization. GSA positioned it as a shared-services platform intended to let agencies evaluate AI without duplicating effort across the government.
OneGov $1-per-agency deals
Over a two-week span GSA announced a series of OneGov agreements making vendor enterprise products available to agencies at $1 per agency per year. The source describes this concentrated procurement period as the largest concentrated federal AI procurement moment on record. The deals announced:
| Date | Vendor | Deal |
|---|---|---|
| 2025-08-06 | OpenAI | ChatGPT for Enterprise at $1/agency/year via MAS |
| 2025-08-07 | AWS | Up to $1B in federal IT savings |
| 2025-08-12 | Anthropic | Claude for Enterprise + Claude for Government (FedRAMP High) at $1/agency/year, all 3 branches |
| 2025-08-13 | Box | Box AI for workflow automation |
| 2025-08-21 | Gemini for Government OneGov |
Anthropic's Claude for Government carries a FedRAMP High authorization, which the source describes as suitable for sensitive unclassified work, and the Anthropic agreement covered all three branches of government. The OpenAI ChatGPT for Enterprise offering was made available through the Multiple Award Schedule (MAS).
The source frames the simultaneous multi-vendor access as a deliberate hedge against vendor lock-in, and reads the $1 price point as vendors treating the government as a strategic distribution channel rather than a revenue source, with embedded usage expected to convert to paid seats or contracts after the one-year term. According to the source, the deals directly implement the procurement pillar of America's AI Action Plan (July 2025). The source also notes that identical pricing across vendors creates a natural experiment in which vendor federal agencies actually prefer, and that the arrangement raises downstream policy questions over whether agency-level AI deployment decisions should be centralized and whether FedRAMP is sufficient for agentic AI that takes actions.
Reception
In an August 14, 2025 response titled "New Federal AI Platform Risks Corporate Capture, Worker Displacement, and Equity Blind Spots," Public Citizen criticized the program on three grounds: corporate capture, arguing that big-vendor dominance is baked in; worker displacement risk; and missing equity and civil-rights safeguards.
Relationships
- supports: America's AI Action Plan.
- deploys-in: Planned GSA — General Services Administration (AI Deployer), DHS — Department of Homeland Security (AI Deployer), DOD — Department of Defense (AI Deployer), FDA — Food and Drug Administration (AI Deployer), ICE — Immigration and Customs Enforcement (AI Deployer).
- deployed-by: Anthropic, OpenAI, planned Amazon, planned Google DeepMind / Google Cloud.
- contradicts: Public Citizen critique on corporate capture.
- related: Sen. Grassley Letter to CISA re ChatGPT (2026-02-05) (security-risk counterpoint), The OpenAI Files (governance concerns), A Framework for AI Development Transparency (Anthropic).