Table of Contents
OpenAI has launched the Grid Protectors Program, a national initiative to make its future AI compute campuses reduce electricity demand when local grids come under strain. The program starts with a planned campus in Effingham County, Georgia, where the company has agreed to provide up to 1,000 megawatts of flexible demand response.
The effort will establish site-specific arrangements with utilities, regulators, and policymakers, setting out how much power each campus can stop drawing, how quickly it can respond, and how long a reduction can last. OpenAI wants that flexibility to become a standard part of how it designs and operates future campuses.
In Georgia, the commitment builds on an existing power agreement; the campus remains in development, but its operating model would give Georgia Power a large, controllable source of demand reduction during periods when homes and businesses need electricity most.
How Campus Flexibility Supports the Grid
Demand response helps balance electricity supply and consumption by reducing customers’ power use when the system is under pressure. An AI campus contributes by lowering the electricity it draws, leaving more of the available supply for other customers.
That is the central mechanism in the Georgia agreement. OpenAI compares the maximum 1,000 MW reduction with the average electricity demand of roughly 700,000 households. The comparison describes the scale of power that could be freed up, rather than electricity the campus would generate.
The national program also includes other potential ways to support a grid. Depending on local needs and available equipment, OpenAI plans to work with operators to supply power from on-site generation or storage. Batteries and power controls can also smooth changes in a campus’s demand.
Those functions address different operating needs. Reducing consumption lowers the load a utility must serve, while supplying stored or locally generated electricity adds a source of power. Smoothing demand changes helps manage fluctuations in the campus’s draw. Each location’s arrangements will determine which capabilities are available.
Georgia’s Pilot Builds on a 25-Year Agreement
Under the 25-year agreement with Georgia Power, the utility can reduce electricity delivered to the Effingham County facility during certain periods to support grid stability. The project is expected to require approximately 3,200 MW of power, with up to 1,000 MW available for flexible demand response.
The flexibility commitment therefore represents roughly 31% of the campus’s planned power requirement. Georgia Power expects the ability to reduce that demand at critical times to lower the amount of new generation it needs to build for future growth.
The campus, known as Project Camellia, is being designed and developed by OpenAI. Its contracted 3.2 gigawatts of power are scheduled to arrive in phases between 2028 and 2032. The Grid Protectors pilot will apply once the campus is operating.
Georgia Power announced in August that the contract had been approved through the regulatory process with the Georgia Public Service Commission. OpenAI has committed to covering the full infrastructure and electric-service costs needed to serve the facility, alongside long-term contractual requirements and financial assurances that protect other customers.
Paying for Capacity Without Shifting Costs
The program’s economic rationale combines steady electricity purchases with the ability to cut demand during peaks. Power lines, substations, and other grid assets carry fixed costs. A campus buying electricity throughout the year can help spread those costs across more sales, while reducing consumption during strained periods can limit the additional capacity needed to cover the busiest hours.
Georgia Power projects approximately $950 million in annual customer savings beginning in 2029 from its portfolio of large-load contracts and expected growth, including the OpenAI project. That is a portfolio-wide projection, not savings attributable solely to Grid Protectors or the Effingham campus. The utility estimates a benefit of at least $15 a month for a typical residential customer using 1,000 kilowatt-hours monthly.
OpenAI’s broader approach also calls for long-term contracts and financial assurances to protect other ratepayers if a campus is delayed, scaled back, or canceled. Those protections address the risk of infrastructure costs remaining after a project’s expected electricity purchases change.
Demand flexibility will accompany investment in new energy supply and transmission, rather than replace it. OpenAI plans to announce additional campus commitments as local arrangements are finalized, with the Georgia agreement as the program’s first model.
