The U.S. General Services Administration (GSA) has officially moved USAi, its government AI testing and deployment platform, from a free pilot to a paid service model. Agencies will now pay a platform fee based on agency size, plus actual AI model usage costs. According to GSA, the goal is to create a sustainable, long-term service as adoption continues to grow, with at least 25 agencies already using the platform.
My take: This move makes sense, but it raises an important debate.
The Case For Charging
- Agencies that use AI should contribute to the cost of maintaining and improving the platform.
- A cost-recovery model can make USAi sustainable rather than dependent on limited government funding.
- Paying for usage may encourage agencies to use AI more strategically and focus on high-value use cases.
- The collected usage data can help GSA negotiate better government-wide AI pricing through initiatives like OneGov.
The Case Against Charging
- Smaller agencies may slow down AI adoption if budgets become a barrier.
- Charging for access could reduce experimentation and innovation, especially during the early stages of AI maturity.
- Agencies that are already struggling with modernization funding may postpone AI initiatives altogether.
- The federal government risks creating a gap between AI-enabled agencies and those that cannot afford to invest at the same pace.
The real question isn’t whether agencies should pay. It’s whether the pricing model will accelerate AI adoption across government or unintentionally slow it down.
As AI becomes a core capability rather than an experiment, sustainable funding is necessary. But success will depend on keeping costs low enough that every agency, regardless of size, can participate in the federal AI transformation.
What do you think? Should government AI platforms be funded centrally to maximize adoption, or should agencies pay for what they use?
Source: Agencies now have to pay to use USAi (Nextgov/FCW, September 3, 2026) [nextgov.com]