The Numbers Come First
For nearly a decade, the conventional wisdom was wrong. Using Federal Energy Regulatory Commission data and U.S. Energy Information Administration retail revenue figures spanning 2015 to 2024, researchers at the Electric Power Research Institute found a causal relationship most consumers never heard about: for every doubling of data center capacity, average retail electricity prices fell by 3.5%. On a statewide basis, that discount widened to roughly 6%.
The mechanism is straightforward economics. Electricity pricing is built on cost recovery, not production cost. More kilowatt-hours consumed means fixed infrastructure costs are divided across a larger base — the same logic that makes a toll road cheaper per driver when traffic rises. Data centers, as massive and predictable load anchors, played that role quietly and effectively for the better part of a decade.
Where the Trend Breaks
The past, however, is not a guarantee. EPRI researcher and study coauthor Asa Watten told Fortune the critical variable going forward is whether AI demand actually materializes at the scale investors are betting on.
'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,' Watten said.
The warning signs are already visible. PJM, the largest power grid operator in the country, projected this week that a $6.3 billion increase in consumer electricity costs over the next three years is mostly attributable to increased data center power demands. In Virginia — home to more data centers than any other state — residential electricity prices have risen more than 13% in the last year, according to EIA data. The broader buildout is expected to reach $7 trillion in spending by 2030.
A YouGov poll of 1,000 Americans found more than two-thirds already expected electricity prices to rise if a data center arrived in their area. Earlier this year Goldman Sachs projected the AI infrastructure buildout would increase electricity costs by 6% between 2026 and 2027, with an additional 3% by 2028.
The Demand Question No One Can Answer
Investor sentiment is wavering. Share prices for Tesla and Alphabet fell this week after both companies announced increases in AI capital expenditures — a signal that markets are beginning to price in the risk of overcapacity rather than reward ambition. Billionaire investor Mark Cuban said on the All-In podcast this week that 'a lot of data centers…are going to be turned into pickleball courts,' arguing that rising power efficiency will reduce the need for raw capacity even as AI adoption grows.
Watten frames the math precisely: if demand falls short, 'your denominator is less than you thought it would be. You're spreading those fixed costs amongst fewer people. It's the opposite of what we want to be doing.'
What Free Markets Require
The EPRI findings deliver a lesson that regulators and politicians eager to restrict data center development should absorb carefully: private capital, when allowed to scale, has historically driven consumer electricity costs down, not up. The threat to that dynamic is not the free market — it is the possibility that trillions in investment are being deployed against demand projections that have not yet been validated by actual customers.
Capital rewards clear rules and honest price signals. If AI adoption meets its projections, the buildout extends a decade-long consumer benefit. If it does not, ratepayers absorb the stranded-cost bill. The productive question for policymakers is not how to throttle data center construction, but how to ensure that the grid's fixed-cost commitments are matched by real, contracted demand — not speculative forecasts dressed up as certainty.
