The Pulse
OpenAI Pledges to Cut Data Center Power During Grid Stress
OpenAI launched its Grid Protectors Program on November 2, 2026, promising to reduce electricity demand at future campuses when local grids face strain. The plan uses a Georgia Power agreement allowing up to 1,000 megawatts of demand respon
AI.info Team · Written with AI · fact-checked ·
“I was in a meeting about a year ago with one of the other large hyperscalers, and they... were defending the five nines, and I don't think anybody's doing that now.”
John McCarrick, OpenAI's head of energy policy
OpenAI is offering a new answer to a growing concern about data-center electricity demand: campuses that can cut their power use when the grid is under pressure. The company announced the Grid Protectors Program on November 2, saying it will work with utilities and regulators to arrange demand reductions at future campuses it develops. The size, speed and duration of those cuts will depend on each location.
Georgia supplies the first 1,000-megawatt model
The program builds on OpenAI’s planned Project Camellia campus in Effingham County, Georgia. Under an agreement with Georgia Power, the facility can reduce demand by up to 1,000 megawatts at times of high need. Georgia Power says the campus is expected to require about 3,200 megawatts and that the utility can reduce the energy it delivers under the 25-year agreement.
OpenAI compares the maximum reduction with the average electricity use of roughly 700,000 households. The campus is not yet operating, and the company says it intends for Effingham to serve as the program’s pilot. The national announcement does not set a uniform reduction target for other sites; OpenAI says it will share further commitments as agreements are finalized.
How a data center could adjust its draw
OpenAI says future arrangements could combine lower electricity use with power from on-site generation or storage, depending on local needs and available equipment. It also points to batteries and power controls as tools for smoothing demand changes. The company has not specified which computing tasks would be reduced or moved, or how much warning a utility would provide before calling on a campus to cut demand.
The pitch rests on two claims about how large computing facilities interact with the grid. OpenAI says campuses can use more electricity when the system has capacity and reduce consumption during peak strain, while steady year-round demand can spread the fixed cost of grid infrastructure across more sales. It argues that this flexibility could limit the need for upgrades built solely to meet peak demand, helping contain costs for other customers.
OpenAI says the Georgia deal could benefit ratepayers
OpenAI says it will pay the full infrastructure and electric-service costs required for its Effingham project. Georgia Power separately says the flexible-demand agreement may allow it to build fewer new generation resources to meet future growth. OpenAI also cites Georgia Power’s projection of about $950 million in annual customer savings beginning in 2029 from its portfolio of large-load agreements and expected growth, including Project Camellia; the utility estimates a benefit of at least $15 a month for a typical home using 1,000 kilowatt-hours.
That savings figure is a portfolio estimate, not a promise that OpenAI’s project alone will reduce bills by that amount. The company also says flexibility cannot replace investment in new electricity supply and transmission. Its proposal is therefore a complement to grid expansion, not a substitute for it—and the central test is whether future campus agreements specify cuts that utilities can call on when the system is strained.