The marriage of data centers and artificial intelligence has captured headlines for its ability to transform modern computing and its monstrous appetite for power.

A new study by Duke University finds that the U.S. power grid can accommodate this growing need for capacity with a relatively simple tool: load flexibility.

The study, “Rethinking Load Growth: Assessing the Potential for Integration of Large Flexible Loads in U.S. Power Systems,” was published by Duke University’s Nicholas Institute for Energy, Environment and Sustainability in February 2025.

The study notes that there is “an increasing urgency” to meet a rising demand for electricity from a variety of factors. The largest of these is the growth of A.I. and data centers, which are expected to account for the “single largest growth segment,” with up to 65 gigawatts (GW) through 2029 and about 44% of U.S. electricity load growth through 2028.

The study offers a solution. Load flexibility can help the grid accommodate that growth, and ironically, A.I. and data centers can facilitate flexibility by working in conjunction with distributed energy resources, such as solar, wind and battery storage.

The Duke researchers define flexibility as “the ability of end-users to temporarily reduce their electricity consumption from the grid” during periods of high demand. They can do this through a number of methods, including on-site generators, shifting workload to other facilities or reducing operations. In fact, many data center operators have stated that they intend to use more on-site generation to help support growing power demands.

When done properly, this flexibility gives planners an added tool. The study notes that when planners can “reliably anticipate the availability of this load flexibility,” expansion pressures decrease and the need for large-scale expansion is lessened.

A little bit of flexibility goes a long way. The study finds that an average annual load curtailment rate of 0.25% will accommodate up to 76 GW of new load, which is equivalent to 10% of the nation’s current aggregate peak demand. The numbers go up from there—with 0.5% of load curtailment freeing up space for 98 GW of new load integration, and 1% curtailment freeing up capacity for 126 GW of new load integration.