Who pays for AI’s electricity? Data centers spark debate over rising power costs
Advocacy groups and community members protest laws surrounding data centers outside the Texas Capitol in Austin on Monday, Feb. 23, 2026.
Austin American statesman/Hearst Newspapers | Hearst Newspapers | Getty Images
Companies racing to build the massive infrastructure needed for the AI boom are facing growing backlash over electricity costs as households and policymakers question whether data centers are driving up electricity bills.
But a recent report from semiconductor research firm SemiAnalytic argued that the expansion of data centers is only part of the story, arguing that market design and policy decisions play a larger role in these energy price increases than the growth of AI infrastructure alone.
From rural Virginia to the Arizona desert, communities that once welcomed technology investment now oppose data centers amid growing concerns that these facilities, called AI hyperscalers, are straining local power grids, increasing costs for everyone else.
Residential electricity prices in the US have increased more than 36% since 2020, from 12.76 cents per kilowatt-hour to 17.44 cents per kilowatt-hour in February 2026, and this price is expected to increase in February. According to data, it reached 19.01 cents per kilowatt-hour as of September 2027. last prediction By the US Energy Information Administration.
“Retail electricity prices have increased faster than the inflation rate since 2022, and we expect them to continue increasing through 2026,” the EIA said in a March 2025 report, ahead of the Iran War.
US President Donald Trump also recently acknowledged the problem in the industry, saying data centers “need some PR help.”
Localized pricing mechanisms
Retail electricity prices in the U.S. reflect the costs of producing, transmitting, and distributing energy, as well as other factors such as taxes and utility investments to improve aging infrastructure.
SemiAnalytics alleged that an obscure market pricing mechanism known as the Base Residual Auction generated most of the “runaway” energy prices in the PJM Interconnection region, a regional grid operator serving 13 eastern states and hosting data centers of hyperscalers such as Google, Anthropic, and Amazon.
Under the mechanism, consumers make payments for expected electricity costs two years in advance, ensuring adequate energy availability during periods of peak demand, such as heat waves or winter storms.
Future energy prices under the mechanism are estimated prices based on expected future demand, calculated through custom models and simulations run on data. However, in all forecasting models, parameters may not always reflect real-world conditions.
SemiAnalysis argued that PJM’s forecasts often overestimate future demand, especially since many planned data centers in the region face construction or installation delays due to chronic memory shortages.
The report compares PJM to another U.S.-regulated energy grid. Electric Reliability Council of TexasIt was stated that prices have remained relatively stable since 2022, despite the development of data center complexes by hyperscalers such as OpenAI, Anthropic and Google.
In the United States, where regulations governing electric grids are decentralized across states and service providers, market design often determines how additional costs are passed on to households.
EIA also noted regional price disparities in a March 2025 report, saying areas with higher residential electricity prices could see increases above the national average.
“In a capacity-constrained market like PJM, prices have risen dramatically as data center demand has increased. But other markets allow for more comprehensive direct cost allocation,” Maeghan Rouch, partner at Bain & Company, told CNBC.
Rouch added that it may not always be entirely clear what is driving increases in consumer energy prices, as unrelated investments in local grids, such as grid strengthening and modernization, or general inflation can also put pressure on households.
“Even without data center investment, we would expect there to be some upward pressure on price growth,” Rouch said.
Commitments from hyperscalers
Large technology companies have also tried to allay concerns about their energy use by making commitments to cover or improve the electricity costs of their own projects. alternative energy sources.
In January, Microsoft made a five-point statement planincluding a commitment to cover any additional electricity costs arising from data centres, among other community investments. This was followed by a similar loyalty By Anthropic in February.
Recently, President Trump summoned executives from leading artificial intelligence companies to the White House. Taxpayer Protection CommitmentEnsuring that expenses incurred by new AI data centers are not passed on to American consumers.
The problem is that the industry isn’t making money, which puts more pressure on them.
Marc Einstein
Research director, Counterpoint Research
Such commitments can be particularly important in “getting support from communities that might otherwise oppose it.” [data center] Especially if data center development is accompanied by alternative investments in local communities, such as jobs or education, according to Chris Howard, head of data centers account management at JLL.
But given that hyperscalers struggle to turn a profit, experts question the legitimacy of such commitments.
“The problem is that the industry isn’t making money, which puts more pressure on them,” said Marc Einstein, research director at Counterpoint Research.
He added that hyperscalers also need to clarify their plans for rising electricity costs. “If they stay quiet about this, it will allow the rumor mill to spiral out of control.”
Technology companies have also committed to meeting data center needs with renewable resources.
According to JLL’s Howard, such alternative energy sources will become increasingly important as concerns about energy availability increase due to increasing demand for data centers around the world.
“The average wait time for grid connection in primary data center markets is currently between four and six years, with up to 10 years in cities like Tokyo,” Howard said.
He added that these energy shortages globally “could create huge opportunities for energy producers, especially when it comes to renewable energy.”
But Howard said skepticism about renewable energy commitments under the current U.S. administration raises questions about how far such sustainability commitments will go in the country.
Even so, analysts say it may be in the corporate interests of AI hyperscalers to honor these commitments.
“It would certainly be better from a public relations standpoint,” Einstein said.
But public outcry could also lead regulators to impose new rules on hyperscalers, Einstein added, “which is not actually what they want.”
