OpenAI Grid Protectors Program to Cut Data Center Power at Peaks

OpenAI Grid Protectors is a new US program, announced on November 2, 2026, under which OpenAI’s future AI compute campuses will cut their electricity use when the local power grid is under strain. The company says it will work with utilities, regulators and policymakers on deals that set how much demand each campus can shed, how quickly and for how long, starting with its planned campus in Effingham County, Georgia.

In its announcement, OpenAI framed the effort as a way to keep electricity “affordable” and “reliable” for families and businesses near its sites, during events such as winter cold snaps or summer heat waves. The arrangements will differ by location, and the company said it will share more commitments as they are finalized.

How OpenAI Grid Protectors is supposed to work

The idea is demand flexibility. Instead of drawing the same amount of power around the clock, a campus is designed to throttle its load when the grid hits peak demand, and to use more power when the grid has room to spare. Depending on local needs, OpenAI says it may also supply power from on-site generation or storage and use batteries and power controls to smooth sudden changes in demand.

OpenAI argues that steady, year-round demand helps spread the fixed cost of power lines and substations across more electricity sales, while flexibility limits the need for expensive upgrades built only for peak hours. “Families and small businesses should not be asked to subsidize a compute campus’s electricity needs,” the company wrote, adding that it has always committed to paying its own energy and infrastructure costs.

Georgia is the pilot

The first test case is Project Camellia, a planned campus in Effingham County, about 45 minutes outside Savannah. According to Georgia Power, the site is expected to need about 3,200 megawatts, delivered in phases between 2028 and 2032, and OpenAI will pay the full infrastructure and electric service costs to serve it.

Under the 25-year agreement, Georgia Power can reduce the power it delivers to the facility by up to 1,000 megawatts during periods of high demand. OpenAI says that is roughly equal to the average electricity use of about 700,000 households, more than Washington, D.C., and Boston combined. The utility calls it one of the largest single-facility demand response commitments in the country.

Georgia Power also projects about $950 million in annual customer savings from 2029 across its portfolio of large-load agreements and expected growth, including Project Camellia, or at least $15 a month for a typical home using 1,000 kilowatt-hours. Those figures are the utility’s projections, and the campus is still in development.

Why it matters

AI data centers are becoming one of the fastest-growing sources of US power demand, and OpenAI cites projections that US data center capacity could exceed 100 gigawatts by 2030. Rising electricity bills near large campuses have turned how much energy AI uses into a local political issue in several states.

The program also lands on a day when the cost of AI infrastructure is in the headlines, with Oracle raising $10 billion in new stock to fund its data center buildout. Governments are paying closer attention to data center capacity as well, as shown by the Russian decree on reserved data center capacity issued the same day.

OpenAI itself notes the limits of the approach: flexibility “cannot replace investment in new energy supply and transmission.” The company has not yet said which other campuses will join the program or what curtailment terms they will carry.