Ahead of the midterm elections, Congress is scrambling to show that it understands voter concerns about AI data centers and their potential impact on electricity rates. But answering the question of how much of those costs will be shouldered by residential utility customers is surprisingly difficult to answer.
That uncertainty is at the heart of a growing fight on Capitol Hill, as lawmakers advance competing proposals aimed at protecting ratepayers in a debate that often belies the complexity of the underlying economics.
On Capitol Hill this week, Republicans have been trying to push through the Ratepayer Protection Act, sponsored by Ohio Republican Sen. Jon Husted. The bill would have state utilities consider holding "large-load" customers responsible for the increased costs.
Husted is in a heated reelection campaign against Democrat Sherrod Brown, who lost his Senate seat in 2024. Brown has made Husted's past support of data centers in Ohio a major campaign issue. The bill easily passed the House in September with strong bipartisan support.
Senate Minority Leader Sen. Chuck Schumer opposes the measure, saying its protections for ratepayers are voluntary. Instead, Schumer is promoting the GRID Savings Act, a similar bill authored by New Mexico Democrat Sen. Martin Heinrich, which he says has more teeth.
Several other data center and AI-related bills are at various stages in the legislative process.
The cost of AI is showing up in customers' bills
The question of how much of the energy and grid costs are being borne by ratepayers is a complicated one and has no universal answer, experts tell NPR. They say it varies by utility and state, but that in the end residential electricity customers will pay for a share of the AI buildout.
By one estimate, Maryland customers, for example, pay somewhere around $168 to $216 more per year on their electricity bills primarily due to the data center boom.
"We're talking tens to hundreds of millions of dollars [of infrastructure] that are being built by these local utilities," says Connor Waldoch, co-founder and chief strategy officer at Grid Status, which focuses on making electricity-market data easier to access and understand. "In many cases, it is just ultimately going to end up in customers' bills."
Tracking the impact of data center growth on residential bills isn't easy, however. How much of the increase is passed on to residential customers and when, depends on utilities' power-purchasing contracts, retail rates and regulatory decisions about who pays, according to Lucy Qiu, a professor at the School of Public Policy at the University of Maryland.
PJM is a regional transmission organization that coordinates the electricity grid and wholesale power market across 13 states and the District of Columbia. It supplies power to Northern Virginia, home to the largest concentration of data centers in the world. PJM's Independent Market Monitor estimates that data centers have cost their 67 million ratepayers about $29 billion over roughly the past two years.
"Everyone pays those costs," says Ari Peskoe, director of the Harvard Electricity Law Initiative. "They're spread across the region to every business and resident that has an electricity meter."
How to determine who the grid infrastructure benefits?
Data center growth can push up wholesale costs of energy by increasing demand. This determines how much generating capacity utilities and regional grid operators need to secure, Qiu says. There are also costs with building new power lines, substations and other infrastructure to connect data centers to the grid. Determining how much of that exclusively serves data centers — and how much also benefits other customers — is part of the difficulty in determining who gets the bill, she says.
"It really depends on the scale of the upgrade and also how utilities are doing the accounting of that different share, whether it's being used solely by the data centers or also being shared by other customers," Qiu says.
No "single, clean number"
Even so, "a single, clean number" that applies nationwide is just not possible, Harvard's Peskoe says, because "a lot of the data about how much particular data centers are paying is hidden behind non-disclosure agreements."
Those NDAs have become a major sore spot on the local level, as residents opposed to new data centers call out local officials for the secrecy that has surrounded many deals. A study last year by the University of Mary Washington found that NDAs had been signed in 80% of Virginia localities where there were existing, approved or proposed data centers.
In an email to NPR, Jeremy L. Slayton, a spokesman for Dominion Energy, which serves Loudoun County, Va., home to about 250 data centers, said, "We recover costs for distribution and transmission infrastructure from customers."
"These assets are core to the reliable electric service we provide our customers. All costs we recover from customers are reviewed and approved by the Virginia State Corporation Commission," he said, referring to the independent state agency that regulates public utilities.
Added costs of building out in a hurry
There is still another factor at play, University of Maryland's Qiu says. The data center industry is in a hurry and waiting for new power-generation technologies doesn't seem to be in the cards. As a result, to meet the demand, utility companies will likely need to opt for less efficient power plants and those costs will be passed onto consumers over time, as well.
One other consideration is optics, Waldoch of Grid Status says. Big customers like AI data centers pay less per kilowatt hour for power than residential customers do. Think of it as a volume discount. In the current environment where 70% of Americans are worried that data centers will increase their electricity bills, Waldoch says, "If you're saying, hey, this data center is actually paying a lot less for electricity than you are, even though it matters much less to them ... it's not like that's palatable for a lot of people at this point."
Benefits of the AI buildout
But not all price increases in recent years are linked to data centers, Qiu says.
"The fuel cost ... is a very important factor," she says and some of those new substations and power lines were going to need to be replaced anyway due to things like extreme weather events.
Qiu also notes that the U.S. power grid needs upgrades to accommodate higher demand for things like electric vehicles. The AI boom "can accelerate the upgrade of the grid infrastructure," she says.
There's also evidence, Qui says, that resiliency — the ability to survive high-impact events such as extreme heat and to recover from outages — is better for people living near data centers due to the robust infrastructure those facilities require. "So if there's a data center, it actually helps … improve the local grid reliability [and] reduce the power outages," says Qui.
Copyright 2026 NPR