On July 14, 2026, PJM Interconnection — the largest grid operator in the United States, serving roughly 67 million people across 13 states and Washington, D.C. — announced the results of its capacity auction for the 2028-29 delivery year. The clearing price hit $325 per megawatt-day, the maximum allowed under its price cap. The auction failed to secure enough future supply to meet its reliability target for the third straight time, falling about 6.8 gigawatts short — equivalent to nearly seven nuclear reactors.
The auction's total cost reached $16.4 billion. According to PJM's independent market monitor, Monitoring Analytics, approximately $6.3 billion — roughly 40 percent — was directly attributable to data center demand. Across the last four auctions combined, data centers have added nearly $30 billion to the capacity costs that all ratepayers absorb.
Moody's Ratings put the mechanism bluntly in a July 2026 report: "the current system lacks adequate mechanisms to ensure that the cost of building new supply is borne by the new entrants and instead socializes new build costs across all customers." When a data center activates hundreds of megawatts of new computing load, the cost of building power plants to serve it does not land on the hyperscaler's bill. It is spread across every household and small business on the grid.
That is the ratepayer fight. And it has become a central issue in the November midterm elections.
The Mechanism
The regulated utility model has worked the same way for a century. When a large new customer wants power, the utility builds the transmission lines, substations, and generation to serve it. Then it recovers those costs across its entire customer base. Ari Peskoe, director of the Electricity Law Initiative at Harvard Law School, summarized it: utilities build, and "we all pay for it."
Two features turn that ordinary mechanism into a wealth transfer.
First, data centers can outbid almost anyone for power. Electricity is only about 20 percent of a data center's total cost base, according to McKinsey, and these are among the most profitable facilities ever built. A hyperscaler's willingness to pay for firm power dwarfs a household's.
Second, residential ratepayers are captive. They cannot shop for a different grid. When data center demand pushes the capacity price to its cap, the AI operator shrugs and a retiree in Ohio does not.
The costs are already showing up on bills. Pepco residential customers in Washington, D.C., saw bills rise about $21 a month starting in June 2025. AEP Ohio customers absorbed roughly a 10 to 15 percent increase over the same capacity year, with commercial customers facing as much as 29 percent.
A separate analysis found that in a 13-state grid territory — including the Senate battlegrounds of Ohio and Pennsylvania — data center demand drove capacity costs up by $9.3 billion, or 174 percent, in a single year.
What States Are Doing
The federal government has not acted decisively, so states have begun to.
Virginia became the first state to assign a meaningful share of grid upgrade costs to data centers. In November 2025, the State Corporation Commission approved a Dominion Energy rate increase that added $16 a month to the typical residential bill but assigned more costs to data center operators to cover necessary grid upgrades. A new GS-5 rate class for data centers with demand of 25 megawatts or more will take effect January 1, 2027. Consumer advocates called it a step forward, though the Piedmont Environmental Council argued it did not go far enough — the decision would still put 61 percent of grid upgrade costs onto individual ratepayers after the 14-year contract period ends.
Oregon took a different approach. Under the state's POWER Act, Portland General Electric will boost rates for data center clients by an average of 29 percent, while residential customers will see their bills drop by an average of 1.3 percent — saving the average ratepayer about $1.91 per month.
Tennessee regulators approved a new wholesale data center rate that will eventually charge data centers roughly 10 percent more for electricity. Nashville Electric Service approved a separate retail rate for data centers and changed how the utility charges them for infrastructure, requiring them to pay more upfront rather than risking costs being passed to other customers.
Wisconsin regulators issued a ruling requiring data centers to cover the full cost of the power infrastructure they require, explicitly stating that utilities "must not shift those costs onto other customers."
Nevada's NV Energy projected that data centers accounted for just 5 percent of its total power sales in 2025. By 2046, they are projected to account for 64 percent. The utility will need to more than double its current electricity load over the next two decades.
States are moving to enforce a beneficiary-pays principle. The entity that creates the demand should pay for the infrastructure that serves it.
The Federal Response
On September 16, 2026, the U.S. House of Representatives passed the Ratepayer Protection Act by a vote of 417-3 — a rare bipartisan margin on an issue that has become politically charged ahead of the midterms.
The bill, sponsored by Rep. Kathy Castor (D-Fla.) and Rep. Gabe Evans (R-Colo.), would require states to consider — but not mandate they adopt — a federal standard ensuring large power customers cover 100 percent of the costs of new generation and transmission upgrades. States would be required to hold a hearing on such a policy within two years of the bill's enactment.
"Large load data centers must cover the full costs of any system updates they require, not families or small businesses," Evans said during floor debate. "We cannot accelerate this growth on the back of Americans working hard to pay their electric bill at the end of the month."
The bill now heads to the Senate, where Ohio Republican Jon Husted has introduced a companion measure. Senate Majority Leader John Thune cautioned that quick passage could prove difficult.
Critics argue the bill is toothless because it does not require states to act. Progressive Rep. Rashida Tlaib (D-Mich.) said it "fails to meaningfully protect our communities" and called for a national moratorium on data centers instead. But the 417-3 vote signaled that both parties have concluded the status quo is politically untenable.
The White House had already attempted a voluntary approach. On March 4, 2026, President Trump established the Ratepayer Protection Pledge, signed by Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and xAI. The pledge committed signatories to "build, bring, or buy all of the energy needed" for their data centers and "cover the cost of all power delivery infrastructure upgrades required." It was expanded in July 2026 to include utilities and state leaders.
But the pledge is voluntary and nonbinding. It commits signatories to a principle without giving regulators a tool to enforce it. As Brookings noted in a July analysis, citing an ICF report, residential electric rates are projected to increase by between 15 and 40 percent by 2030 absent new tariff structures — and rates could double for some utilities by 2050.
The China Difference
China faces the same physics. It does not face the same political mechanism.
China's residential electricity price averaged approximately 0.5 yuan per kilowatt-hour in 2025, roughly 42 percent of the U.S. average of 1.2 yuan per kilowatt-hour, adjusted for exchange rates. This reflects a regulatory logic in which residential rates are set by the state and subsidized by higher industrial and commercial rates — a policy described as "people's livelihood first."
In China's western regions, industrial electricity prices are among the lowest in the world. In Yunnan, western Inner Mongolia, Xinjiang, Gansu, and Ningxia, average industrial rates ranged from 0.35 to 0.42 yuan per kilowatt-hour in 2025, about 70 percent of the national average. These regions generate power primarily from hydropower, wind, and solar.
The "East Data, West Computing" strategy channels computing demand to these low-cost regions. In Inner Mongolia, Gansu, and Ningxia, average industrial rates are approximately 0.41 yuan per kilowatt-hour, compared with roughly 0.60 yuan per kilowatt-hour in the U.S.
When a data center is built in Inner Mongolia, it does not bid up the capacity price for residents in Beijing or Shanghai. It connects to a grid where the power is already there, the land is cheap, and the climate reduces cooling loads. The cost of serving the data center is borne by the data center, because the pricing structure does not socialize it across residential ratepayers.
That is not a story about virtue. It is a story about governance. China's system is designed to route computation to where the resources are. The U.S. system is designed to let utilities recover infrastructure costs across their entire customer base — a model built for a world where large new loads arrived slowly and were shared broadly.
What the Ratepayer Fight Reveals
The ratepayer issue is less a technical problem than a structural one. Closed-loop cooling, behind-the-meter generation, separate rate classes for large loads, and beneficiary-pays tariffs are all available. The challenge is the utility model itself, which was designed for a world of incremental demand growth where new infrastructure costs could be spread across millions of customers without materially affecting any individual bill.
AI demand is not incremental. It arrives in gigawatt-scale blocks, in 18-month windows, in communities that did not ask for it. When costs are socialized, political backlash tends to follow. And when the backlash becomes an election issue — as it has in Pennsylvania, Ohio, and at least 18 other states where candidates are running television ads mentioning data centers — the pressure to act becomes overwhelming.
The House vote of 417-3 was a signal. The states that have already acted — Virginia, Oregon, Tennessee, Wisconsin — are a preview. Data centers are increasingly being asked to pay their own way. The remaining states are likely to follow.
Sources:Fortune (July 24, 2026); Insurance Journal (July 15, 2026); Business Model Analyst (July 29, 2026); Inside Climate News (January 7, 2026); Columbia Daily Herald (August 27, 2026); Politico (September 16, 2026); The Hill (September 15, 2026); BSS/AFP (September 17, 2026); Utility Dive (November 13, 2025); Yahoo Finance (April 29, 2026); Brookings (July 9, 2026); IndexBox (July 8, 2026); USA Today (July 7, 2026); China Times (August 18, 2026).
Disclaimer
The information provided in this article is for general informational and educational purposes only. It does not constitute legal, financial, or professional advice. The author and publisher are not responsible for any actions taken based on the content of this article. Readers should consult qualified professionals for advice specific to their situation. All trademarks and references to third-party products, services, or organizations are the property of their respective owners. The performance data and benchmarks discussed are based on specific research studies and may not generalize to all use cases or environments. As of the publication date, the AI landscape continues to evolve rapidly, and readers should verify current information independently.
Limitations
This analysis is based on reporting and public data available as of the article date; figures may be revised as sources update.
Forecasts from third-party analysts can change with market conditions.
Cost and pricing examples are point-in-time estimates; actual rates vary.
Country and company comparisons rely on public reporting, not operational data.
This sector moves fast; timelines and deal terms may be updated later.
Company deals and regulatory rulings may evolve; verify current status.
AI infrastructure is changing quickly; claims can become outdated soon.
Sources
- Fortune (July 24, 2026)
- Insurance Journal (July 15, 2026)
- Business Model Analyst (July 29, 2026)
- Inside Climate News (January 7, 2026)
- Columbia Daily Herald (August 27, 2026)
- Politico (September 16, 2026)
- The Hill (September 15, 2026)
- BSS/AFP (September 17, 2026)
- Utility Dive (November 13, 2025)
- Yahoo Finance (April 29, 2026)
- Brookings (July 9, 2026)
- IndexBox (July 8, 2026)
- USA Today (July 7, 2026)
- China Times (August 18, 2026).
The information provided in this article is for general informational and educational purposes only. It does not constitute legal, financial, or professional advice. The author and publisher are not responsible for any actions taken based on the content of this article. Readers should consult qualified professionals for advice specific to their situation. All trademarks and references to third-party products, services, or organizations are the property of their respective owners. The performance data and benchmarks discussed are based on specific research studies and may not generalize to all use cases or environments. As of the publication date, the AI landscape continues to evolve rapidly, and readers should verify current information independently.
Limitations: This analysis is based on reporting and public data available as of the article date; figures may be revised as sources update.; Forecasts from third-party analysts can change with market conditions.; Cost and pricing examples are point-in-time estimates; actual rates vary.; Country and company comparisons rely on public reporting, not operational data.; This sector moves fast; timelines and deal terms may be updated later.; Company deals and regulatory rulings may evolve; verify current status.; AI infrastructure is changing quickly; claims can become outdated soon.