OpenAI's Grid Protectors Program promises data centres that power down when the grid is strained

The first test is OpenAI's planned campus in Effingham County, Georgia, which is still in development[1]. Georgia Power says the facility is expected to need about 3,200 MW and that, under a 25-year agreement, the utility can reduce energy delivered to it during certain periods; OpenAI has committed up to 1,000 MW of that as flexible demand response, which Georgia Power calls among the largest single-facility commitments of its kind in the US[2]. OpenAI puts 1,000 MW at roughly the average demand of 700,000 households[1].

OpenAI says it will pay the full infrastructure and electric-service costs needed to serve the site, and that depending on local needs it may also supply power from on-site generation or storage and use batteries and power controls to smooth changes in demand[1].

The pitch to ratepayers is that grid infrastructure is a fixed cost. A campus that draws steadily when there is spare capacity and backs off at peaks helps pay for that infrastructure without forcing upgrades sized for the worst day[1]. Georgia Power's contract with OpenAI has been approved by the state Public Service Commission, and the utility projects about $950 million a year in customer savings from 2029 across its whole portfolio of large-load contracts and expected growth, or at least $15 a month for a typical home using 1,000 kWh[3].

Power is now the bottleneck for AI build-outs, and households worried about their bills are a political risk for every lab building campuses. Grid Protectors is OpenAI's attempt to get ahead of that by turning its biggest liability, enormous new demand, into a resource utilities can call on. OpenAI itself concedes the limit: flexibility cannot replace investment in new generation and transmission[1]. Beyond Georgia, the program has no published commitments yet, so the real test is whether the next campuses sign similar terms.