AI data center ‘crazy’ is driving up electricity bills – here’s why

Energy experts say data centers powering the artificial intelligence revolution are driving up household electricity bills, and relief may not come soon.

Retail residential electricity prices rose 7.4% in September to about 18 cents per kilowatt-hour, according to the latest data from the Energy Information Administration.

EIA’s May forecast showed that electricity prices roughly tracked inflation from 2013 to 2023, but are likely to outpace inflation until at least 2026. Some regions will be hit harder than others, the newspaper said.

Energy experts and economists point to the power-hungry data centers that support AI projects as a key reason for the soaring prices.

These data centers are vast warehouses housing computer servers and other IT equipment that power cloud computing, artificial intelligence, and other technology applications.

The basic reason for the price increase: Electricity demand, including actual and projected demand, exceeds new supply.

Data centers are expected to consume 6.7% to 12% of total U.S. electricity by 2028, up from 4.4% in 2023, the U.S. Department of Energy estimated in December 2024.

John Quigley, a senior fellow at the Kleinman Center for Energy Policy at the University of Pennsylvania, pointed to the “data center craze” as a major factor in the rise in household electricity prices.

“When it comes to electricity demand growth, those are pretty much everything,” Quigley said.

“The situation will get even worse,” he said.

To be sure, data centers are not the only culprit behind rising electricity prices, experts say.

But rising electricity prices “could strain household budgets, undermine economic competitiveness, and impede the electrification of the energy system,” researchers at Lawrence Berkeley National Laboratory wrote in a recent analysis.

The rise in electricity prices for American households also comes as politicians continue to use affordability themes to drum up support.

New Jersey Governor-elect Mikie Sherrill and Virginia Governor-elect Abigail Spanberger, both Democrats, have promised to lower electricity bills for residents in their states. “We’ll make sure data centers don’t drive up energy costs for other Virginians,” Spanberger said during his campaign.

During his campaign, President Donald Trump also promised to cut electricity and energy prices in half within his first 18 months in office.

“Affordability remains the most salient issue in politics,” Washington Research Group strategist Chris Krueger said in a research note Tuesday.

Rising utility costs are pushing households deeper into debt, according to a recent analysis by the Century Foundation, a progressive think tank.

The average past due utility balance has increased 32% since 2022, from $597 to $789, according to the study. Utility charges include electricity and other costs such as gas and water.

Winter heating costs for households that use electricity for heating are estimated to rise to $1,205 this season, an increase of about 10% from last winter’s $1,093, according to the National Energy Assistance Directors Association.

“Consumers may again feel pressure on utility bills in the coming months, especially if we have a cold winter,” according to an October report from the Bank of America Research Institute.

AI euphoria is driving the U.S. stock market higher than ever, fueling speculation that the market is in a tech-driven bubble that will burst soon.

Regardless of whether the rise in the AI ​​market is sustainable, the scale of the technology’s growth is undeniable. The International Energy Agency predicts that global electricity demand from AI data centers will more than quadruple by 2030.

“Global electricity demand from data centers will more than double over the next five years, and by 2030 will consume as much electricity as Japan as a whole consumes today,” IEA Director-General Fatih Birol said in the analysis.

The impact will be “particularly strong” in countries like the United States, where data centers are expected to account for almost half of overall electricity demand growth, according to the IEA analysis.

In 2030, the U.S. economy will use more electricity to process data than to manufacture all energy-intensive products, including aluminum, steel, cement and chemicals, combined, according to an IEA study.

Projected demand is increasing the need for new infrastructure such as transmission lines, substations and power plants, and companies are passing on at least some of that cost to residential consumers, UPenn’s Quigley said.

In other words, households are partially subsidizing the expansion of AI data centers, he said.

“AI-driven power demand is occurring across the country, but some grid managers are better able to control costs than others,” Quigley said.

“The amount of increase (in price) will vary by region,” he said.

For example, an October report from Lawrence Berkeley National Laboratory, a U.S. Department of Energy laboratory managed by the University of California, found that extreme weather events such as hurricanes, storms and wildfires contributed to “significant” price increases in some states such as California, where wildfire risk mitigation and liability insurance were “key cost drivers.”

Researchers at Lawrence Berkeley National Laboratory found that electricity prices actually fell in 31 states from 2019 to 2024, after accounting for the effects of inflation. Seventeen states saw postinflationary price increases, particularly on the West Coast and Northeastern states.

Nationally, average retail electricity prices rose 23% in that period in nominal terms, before accounting for inflation, their study found.

According to the Bank of America Research Institute, the increase in household electrification, including electric vehicles, is one of the factors driving up electricity demand.