Rockefeller and OpenAI foundations launch a $50 million city-services initiative

Funding targets a defined public-service problem

The Rockefeller Foundation's announcement launches Cities Deliver with OpenAI Foundation. The $50 million initiative will support eligible nonprofit and civil-society organizations working with city governments on AI-enabled public services.

The call opened October 14 and closes November 29, 2026. The announcement lists grants from $500,000 to $2 million and an ambition to reach up to 100 million city residents globally. That reach is a program aim, not a result already achieved.

Eligibility includes a government partner

The four focus areas are public health, economic mobility, environmental resilience and civic engagement. Each project must have an eligible lead organization, a committed city government partner and a charitable purpose.

Projects need a meaningful frontier-AI component for complex tasks, and may combine it with existing city tools and systems. Applicants can choose the models and products that suit their needs; the announcement specifies no requirement or preference to use OpenAI technology.

Civic-engagement proposals exclude voting and election projects. Organizations should read the full current requirements through the announcement before treating a broadly relevant idea as eligible. Advancing applicants will be invited to a further selection stage.

The program builds on earlier city collaborations, with planned expansions involving flood alerts, dengue forecasting and resident engagement. Those examples show the kinds of service problems being considered, without guaranteeing that a new city's implementation will reproduce their outcomes.

Define how residents will benefit

Franklin analysis: start a proposal with one service bottleneck and a city partner able to measure it. Specify who will use the result, what decision it supports and how residents can obtain help if the AI-assisted process fails.

Choose a baseline such as response time or access to an existing service, then define a measurable improvement. A clear delivery and evaluation plan is more useful than adding a model to a proposal whose public benefit remains unspecified.