As hyperscale computing expands, public opposition to data centers has become closely tied to concerns that utility bills will rise. A YouGov poll conducted last year found that, among 1,000 Americans surveyed, more than two-thirds expected electricity prices to increase if a data center were built in their area. Earlier this year, Goldman Sachs projected that the AI infrastructure buildout would raise electricity costs by 6% between 2026 and 2027, followed by an additional 3% increase by 2028.

A recent working paper from the Electric Power Research Institute, however, complicates the assumed relationship between the AI boom and household electricity bills. The research found that through at least 2024, data center operations had run counter to consumer concerns and were associated with lower retail electricity costs. Using data from the Federal Energy Regulatory Commission (FERC) and retail revenue data from the U.S. Energy Information Administration from 2015 to 2024, researchers identified a causal relationship between data center demand and electricity prices: for every doubling of data center capacity, average retail electricity prices fell by 3.5%. At the state level, the decline was about 6%.

Much of the relationship can be explained by economies of scale, a key issue because utilities must recover the cost of generation, transmission, and distribution infrastructure through customer bills.

“Electricity markets are different than a lot of markets that they interact with,” Asa Watten, the study’s co-author and an EPRI researcher, told Fortune.

Unlike markets for soybeans or gasoline, where prices are tied to the cost of production, electricity prices are based on cost recovery, or the amount of electricity consumed. When fixed costs are spread across more consumers and more kilowatt hours—the standard unit of energy—the greater volume of kilowatt hours divides those fixed costs more widely, lowering prices. Increased data center usage also raises load, bringing more generators online, many of which are becoming more energy efficient.

That pattern, however, is not guaranteed to continue. If it reverses, the shift could point to a broader challenge for the future of AI infrastructure: whether the grid is expanded at a pace that matches real electricity use. PJM, the largest power grid operator in the United States, projected in a report this week that a $6.3 billion increase in consumer electricity costs over the next three years can be attributed mostly to rising data center power demand. The growth in data center construction, expected to reach $7 trillion in spending by 2030, is already correlated with higher power costs. In Virginia, the state with the most data centers, residential electricity prices have increased by more than 13% over the past year, according to data from the U.S. Energy Information Administration.

What will determine the future relationship between data centers and electricity costs?

Watten said the biggest determinant of future electricity prices will be whether the rapid AI buildout produces the level of demand that developers and utilities expect.

“If the grid builds capacity, expecting a lot of demand from data centers, and that doesn’t show up, that could be a clear story of how data centers could increase prices in the future in a way that they did not do in the past,” he said.

Data centers are expected to carry substantial fixed costs. If customers for those facilities do not materialize, “then your denominator is less than you thought it would be,” Watten continued. “You’re spreading those fixed costs amongst fewer people. It’s the opposite of what we want to be doing, so that could increase prices.”

The discussion comes amid broader debate over whether AI is in a bubble and when such a bubble might burst. Some signs suggest investors are becoming more skeptical of the technology’s near-term payoff. On Thursday, shares of Tesla and Alphabet fell after both companies announced increases in AI capital expenditures.

In an episode of the All-In podcast this week, billionaire investor Mark Cuban warned that “a lot of data centers…are going to be turned into pickleball courts.” Cuban said hyperscalers are right to assume AI adoption will keep increasing, but argued that AI will become cheaper to use as power efficiency improves, meaning that all of the data center capacity now being built may not be needed.

Watten said there is also a more optimistic interpretation. He said he does not like to speculate about the future of AI or what it means for the data center buildout, but added that, in general, energy will continue to become more efficient. Electrification—including more electric vehicles and electric heat pumps, in addition to data center growth—could continue to reduce household energy costs in a way that does not depend on an AI boom.

“This clearly efficiency-increasing thing or total budget-reducing thing could have positive spillovers to your neighbors,” Watten said, “such that more electric cars means that if done well, prices are also going down—or at least not going up.”

This story was originally featured on Fortune.com: https://fortune.com/2026/07/26/data-centers-electricity-costs-cheaper-7billion-buildout-ai-demand/