Who pays for AI’s electricity? Data centers spark debate over rising power
Advocacy groups and community members protest laws surrounding data centers while outside the Texas Capitol in Austin Monday, Feb. 23, 2026.
Austin American-statesman/hearst Newspapers | Hearst Newspapers | Getty Images
The companies racing to build the massive infrastructure needed for the artificial intelligence boom are facing growing backlash over electricity costs, as households and policymakers question whether data centers are driving up power bills.
However, a recent report from SemiAnalysis, a semiconductor research firm, argued that the expansion of data centers is only part of the story, and claimed that market design and policy decisions play a greater role in these energy price increases than AI infrastructure growth alone.
From rural Virginia to the Arizona desert, communities that once welcomed tech investment are now pushing back against data centers amid growing concerns that these facilities — built by so-called AI hyperscalers — are straining local power grids, raising costs for everyone else.
Since 2020, residential electricity prices in the U.S. have risen by more than 36%, from 12.76 cents per kilowatt-hour to 17.44 cents per kilowatt-hour in February 2026, and are expected to hit 19.01 cents per kilowatt-hour by September 2027, according to the latest forecast by the U.S. Energy Information Administration.
“Retail electricity prices have increased faster than the rate of inflation since 2022, and we expect them to continue increasing through 2026,” the EIA said in a March 2025 report before the Iran War.
U.S. President Donald Trump recently also acknowledged the problem for the industry, saying data centers “need some PR help.”
Localized pricing mechanisms
Retail electricity prices in the U.S. reflect the costs of generating, transmitting, and delivering power, along with other factors such as taxes and utility investments to upgrade aging infrastructure.
SemiAnalysis claimed that an obscure market pricing mechanism known as the Base Residual Auction accounted for most of the “runaway” energy prices in the PJM Interconnection area — a regional grid operator serving 13 eastern states and home to data centers from hyperscalers like Google, Anthropic, and Amazon.
Under the mechanism, consumers make payments for expected electricity costs two years in advance, ensuring sufficient power availability during peak demand periods, such as heatwaves or winter storms.
Future energy prices under the mechanism are forecasted prices based on anticipated future demand, calculated through simulations run on proprietary models and data. But with all prediction models, parameters may not always reflect real-world circumstances.
SemiAnalysis argued that PJM’s forecasts often overestimated future demand, particularly as many planned data centers in the area faced construction or assembly delays due to a chronic memory shortage.
The report contrasted PJM with another energy grid overseen by the Electric Reliability Council of Texas, where…
Read More: Who pays for AI’s electricity? Data centers spark debate over rising power