For years, the conventional wisdom has been that data centers are a drain on the power grid, driving up electricity bills for everyone. But a new analysis from the Electric Power Research Institute (EPRI) complicates that narrative, showing that up until 2024, the opposite was true: data centers actually helped lower retail electricity prices. The research, which examined data from the Federal Energy Regulatory Commission and the U.S. Energy Information Administration between 2015 and 2024, found a clear causal link. For every doubling of data center capacity, average retail electricity prices fell by 3.5%. On a statewide level, the reduction was even steeper at about 6%.

How does that work? Asa Watten, the study's coauthor and an EPRI researcher, explains that electricity markets operate differently from commodity markets like soybeans or gasoline. In those markets, prices are driven by production costs. But electricity prices are based on cost recovery — how much power is consumed. As fixed costs are spread across more kilowatt-hours, the per-unit cost drops. Data centers, by consuming vast amounts of electricity, help distribute those fixed costs over a larger base. Additionally, their load increases often bring more efficient generators online, further lowering costs.

But this trend may be short-lived. The massive buildout of data centers to support artificial intelligence — projected to reach $7 trillion in spending by 2030 — is already showing signs of reversing the downward pressure on prices. PJM Interconnection, the largest power grid operator in the United States, released a report this week estimating that $6.3 billion in rising consumer electricity costs over the next three years can be attributed largely to data center demand. In Virginia, the state with the highest concentration of data centers, residential electricity prices have jumped more than 13% in the past year, according to the EIA.

Watten points out that the critical factor is whether the anticipated AI demand actually materializes. If utilities build capacity expecting a surge in load from data centers, and that demand fails to appear, the fixed costs of that infrastructure would be spread among fewer customers, driving prices up. "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. He adds that if the denominator — the number of customers — is smaller than expected, "you're spreading those fixed costs amongst fewer people. It's the opposite of what we want to be doing."

There are already signs that investors are growing wary. On Thursday, shares of Tesla and Alphabet fell after both companies announced increases in AI capital expenditures. On the "All-In" podcast, billionaire investor Mark Cuban warned that many data centers may end up as "pickleball courts" — a metaphor for underutilized assets. His reasoning: hyperscalers are betting on continued AI adoption, but as AI becomes more energy-efficient, the massive capacity being built might not be needed.

Yet there is a more optimistic scenario, Watten notes. Electrification — through electric vehicles, heat pumps, and data centers — could continue to drive down household energy costs if managed well. Efficiency gains can have positive spillover effects. "More electric cars means that if done well, prices are also going down — or at least not going up," he said. The fate of electricity prices now hinges on whether the AI boom delivers on its promise or leaves behind a costly oversupply.