Greg Robinson, cofounder and CEO of Aston, writes for the Forbes Technology Council about two studies that illustrate on-grid and off-grid approaches to the future of U.S. power. The first study, by the Nicholas Institute, focuses on creative ways to become flexible with the grid to manage the spike in energy demand due to the rise of AI data centers and the electrification of everything. The other study focuses on off-grid solar microgrids as a way to solve AI energy spikes.

The International Energy Agency recommended that data centers be built in areas with more grid capacity and more power for their energy usage. A recent Nicholas Institute study suggested that data centers could operate more flexibly, temporarily reducing their energy consumption from the grid during times of peak demand.

John Quigley, a senior fellow at the University of Pennsylvania’s Kleinman Center for Energy Policy, tells Floodlight that plenty can be done to update the existing grid’s capacity to meet the electricity needs of data centers without immediately installing new transmission lines. He pointed to a recent Nicholas Institute study that finds load flexibility could be a “key solution to the United States' soaring electrical demand.”

U.S. power demand has soared in recent years with growing artificial intelligence, construction and electrification needs. The Bloomberg Energy Daily cites new Nicholas Institute research saying the United States can meet this rising consumption and quickly add huge loads to its grids without building new power plants by deploying grid flexibility—strategically cutting consumption by the equivalent of about one day a year’s usage.

During a U.S. House Subcommittee on Energy hearing, Duke University expert Tyler Norris discussed a recent study he coauthored on how to meet rising energy demand in the United States. "Our findings suggest that with modest flexibility from new large electricity customers, the existing U.S. power system can accommodate substantial load additions without compromising reliability," Norris testified, according to POWER magazine.

A new paper from Harvard University argues that the cost of upgrading power systems to provide electricity to data centers will fall to normal people and businesses under traditional utility models, Heatmap News reports. In proposing alternative approaches, the Harvard researchers cite a recent study from Duke University scholars that shows utilities could avoid billions of dollars of system upgrades if data centers commit to powering down for a small portion of every year.

As utilities face a surge in electricity demand, a study by Duke University scholars offers an alternative to costly new generation: Let data centers come online under an agreement that they won’t be able to draw power from the grid when energy use is at its highest levels. “If you’re not putting more stress on some of the existing peaks, you may be able to defer that capacity addition and come online more quickly,” lead author Tyler Norris told The News & Observer.

On March 5, testifying before the U.S. House Energy and Commerce Committee’s Subcommittee on Energy, Tyler Norris cited a recent Nicholas Institute study on how load flexibility could help the existing U.S. power grid meet elevated demand.

A recent Duke University study found that electric utilities can already handle demand growth from data centers for most of the year. They just need to shift when those power requests happen during a couple of hours in the year. “If these loads were able to reduce consumption at critical times, then we could accommodate them without triggering the need for this new generation or transmission,” coauthor Dalia Patiño-Echeverri, associate professor at the Duke Nicholas School of the Environment, told WFAE. 

Canary Media reports that if data centers and other big electricity customers limit their power use during peak hours, it could unlock tens of gigawatts of ​“spare” capacity on US grids, citing a recent analysis from Duke University’s Nicholas Institute.