The Hidden Cost of the AI Boom: Why Your Electricity Bill Might Rise

Technology

The Hidden Cost of the AI Boom: Why Your Electricity Bill Might Rise

By reviewsideal Staff  |  September 27, 2026  |  8 min read

The AI boom is usually described in terms of chips and models. Its most tangible cost for ordinary people may show up somewhere less glamorous: the monthly power bill. In parts of the United States, the grid's tightest market is already charging more because of data center demand, and households are paying part of the difference.

Scope note: the clearest evidence comes from the U.S. mid-Atlantic and Texas grids. Electricity markets work very differently across countries and even across U.S. states, so what happens on your bill depends on where you live and who regulates your utility.

How a Data Center Ends Up on Your Bill

AI does not add a line item to your bill. The effect travels through the system. Grid operators must guarantee enough generating capacity for future peak demand, and when large new data centers are forecast to add load, the price of that guarantee rises. Utilities pass those costs on to customers through retail rates. On top of that come the transmission lines and substations needed to connect huge new loads, and the cost of those upgrades is often spread across all customers in the region.

The Numbers From the Biggest U.S. Grid

PJM Interconnection runs the largest U.S. wholesale power market, serving about 67 million people across 13 states and Washington, D.C. Its annual capacity auction is where the effect is easiest to see. Prices there rose from about $29 per megawatt-day for 2024/25 to about $270 for 2025/26 and about $329 for 2026/27. The market's independent monitor estimated that data centers were responsible for 63% of the price increase in the 2025/26 auction, worth roughly $9.3 billion in extra costs recovered from customers in a single year.

Some bills have already moved. Pepco residential customers in Washington, D.C., saw an average increase of about $21 a month starting in June 2025, and New Jersey customers reported jumps in the high teens to about 20% that same month. Wholesale costs kept climbing in 2026: reporting in May said PJM's first-quarter wholesale prices were up 76% from a year earlier, driven largely by capacity costs. PJM itself has said the latest auction results would lift some customer bills by roughly 1.5% to 5%, depending on the utility and state, while other analysts project much larger cumulative increases later in the decade.

The Price Path in PJM's Capacity Auctions

  • 2024/25: about $29 per megawatt-day
  • 2025/26: about $270, a roughly ninefold jump
  • 2026/27: about $329, another increase of about 22%
  • After temporary caps end: environmental group NRDC estimates costs could climb by $27–30 billion a year from 2028 to 2032, about $163 billion in total, if no further action is taken

One widely quoted estimate puts the average household in the PJM region facing around $70 more per month by 2028. Treat that as a projection with wide error bars, since it depends on caps, new generation, and how many planned data centers are actually built.

Texas Is Bracing Too

Texas runs its own grid, ERCOT, which set a record peak of about 91,300 megawatts in July 2026, with planning estimates for 2030 far higher. A Texas research group has projected that the average household could pay about $580 more per year in 2030 than in 2025, again a projection rather than a certainty. The state has responded with Senate Bill 6, which lets the grid operator disconnect large non-critical loads during emergencies and encourages big users to cut back or switch to on-site generation when supply is tight.

The Other Side of the Argument

Attributing bills to data centers is not simple. Rates are also rising for other reasons: aging grid equipment, extreme weather, natural gas prices, and general inflation. Much of the pressure comes from forecasts of demand, and those forecasts can prove too high if projects are delayed, canceled, or built more efficiently than expected. Utilities and the tech industry also argue that large, steady customers can help spread fixed grid costs over more users, and that data centers should pay their own way through special rates. Researchers and regulators disagree on how much cost has truly been shifted onto households, and the answer differs by state.

What Regulators and Tech Companies Are Doing

  • Price caps and emergency auctions: temporary limits on capacity prices and a federal-and-state effort to buy new supply for the PJM region.
  • Large-load tariffs: special rate classes that make big data centers commit to paying for the power and grid upgrades they trigger.
  • Bring your own power: proposals that require new data centers to build or contract their own generation, a strategy behind the nuclear and gas deals tech companies have been signing.
  • Demand flexibility: agreements to curtail data center use when the grid is stressed.
  • Pauses and rollbacks: some states have debated moratoriums, including a proposed multi-year pause on new data centers in New York, and tax-incentive reviews.

What You Can Do

You cannot control wholesale markets, but you can limit your exposure. If you live in a state with retail choice, compare supplier plans and understand when your fixed rate ends. Ask your utility about time-of-use rates and demand-response programs that pay you to shift usage away from peak hours. Efficiency upgrades, from insulation to smart thermostats, cut your kilowatt-hours regardless of price. And because rate cases and data center approvals are decided by public utility commissions, comment periods and hearings are one of the few places ordinary customers have direct influence.

The Bottom Line

The AI boom does not automatically mean higher bills for everyone, but in the tightest grids it already has, and the projections suggest more pressure ahead unless new supply arrives quickly or data centers are made to cover their own costs. The real question is who pays for the build-out: the companies that profit from AI, or everyone connected to the same grid. That is a policy choice, and it is being made right now, one rate case and one auction at a time.

Follow reviewsideal for ongoing coverage of what the AI build-out means for energy and costs.

Post a Comment

Previous Post Next Post