Families could get see higher electric bills if AI data boom goes bust

Houses near a data center in Ashburn, Virginia, USA, on Friday, July 25, 2025.
Bloomberg | Bloomberg | Getty Images
Data centers that are not yet built are driving up electricity prices and could leave consumers dependent on expensive energy infrastructure if demand forecasts are inaccurate.
The race to build AI-enabled facilities has fueled a boom in data centers that train and run large language models like OpenAI’s ChatGPT and Anthropic’s Claude, upending a utility industry accustomed to no increase in electricity demand for 20 years.
But now some investors and energy market analysts are questioning whether the AI race has turned into a bubble; a bubble that will be expensive to unravel as new transmission lines and power plants are built to support these data centers.
Consumers served by the largest electric grid in the U.S. will pay $16.6 billion to secure future power supplies to meet demand from data centers from 2025 to 2027. observer report It was published this month.
Grill PJM InterconnectionThe world’s largest data center hub, serving more than 65 million people in 13 states, is located in Virginia and fast-growing markets such as Northern Illinois and Ohio.
About 90% of that bill, or $15 billion, will be used to pay for future data center demand. Monitoring AnalyticsPJM’s independent market monitor. This represents a “tremendous transfer of wealth” from consumers to the data center industry, he told watchdog PJM. 10 November letter.
“Many of us are very concerned that we will be paying today for a data center tomorrow,” said Abe Silverman, general counsel for the board of utilities in New Jersey, one of the states PJM served from 2019 to 2023. “It’s a little scary unless you really trust the load forecast.”
Residential electricity prices in September rose 20% in Illinois, 12% in Ohio and 9% in Virginia compared to the same period last year, according to federal data. Energy Information Management. Each of these states ranks among the top five markets for data centers in the US
Joe Bowring, president of Monitoring Analytics, said the costs associated with securing power for data centers are reflected directly in consumer electricity bills. “When wholesale energy costs go up, people pay more, and when they go down, they pay less,” he said.
Forecast uncertainty
PJM predicts 30 gigawatts of energy extra demand from data centers by 2030, but it’s unclear how much will eventually materialize. This is equivalent to the average annual energy consumption of more than 24 million homes in the United States.
Cathy Kunkel, a consultant at the Institute for Energy Economics and Financial Analysis (IEEFA), said data center developers research projects in different locations before committing to a site, so there are likely duplications in the estimates.

“We’re in a bit of a bubble,” said Silverman, the New Jersey official. “There’s no doubt that data center developers are getting out of hand by making a lot of new demands. It’s impossible to say exactly how many of these are speculative or real.”
Independent energy producers Constellation EnergyThe largest owner of nuclear power plants in the USA and Vista Corp. It warned earlier this year that data center demand forecasts were likely inflated.
“I have to tell you guys, I think the burden is overrated. We need to put the brakes on here,” Constellation CEO Joe Dominguez said in the company’s May earnings report.
Meanwhile, Vista CEO James Burke also said in May that data center demand could be overestimated by three to five times in some regions as developers review their projects across the country.
‘Stable cost’
The risk is that utilities invest in expensive infrastructure to meet data center demand, but not all of those facilities end up being built or consume less electricity than expected, consultant Kunkel said.
“Consumers — residential, commercial and other industrial ratepayers — are often the ones paying for overbuilt electrical infrastructure,” Kunkel said. A potential problem would arise if capacity that was not needed was created, which could “leave taxpayers on the hook for costs.”
Data center demand forecasts fell as utilities began implementing stricter rules.
For example, in Ohio, American Electric Power Recently, requests for 30 gigawatt electrical connections came from data centers.
AEP proposed stricter rules “to reduce the risk of transmission infrastructure being built for speculative data center projects,” according to a filing filed with the state public utility commission in May 2024.

AEP rules require data centers to pay for 85% of the energy they claim to need, even if they actually use less to cover infrastructure costs. It also implemented an exit fee if data centers canceled their projects or failed to meet contractual terms.
AEP’s data center requests in Ohio fell by more than half to 13 gigawatts after the public utility commission approved the rules last July.
“When faced with potential financial commitments, most speculative or uncertain data center projects did not submit load work requests as intended,” the Columbus, Ohio-based utility said in a statement.
It was stated that the number of requests may decrease further as the new rules force data centers to enter into binding contracts.
The Data Center Coalition, a lobbying group for big tech companies, and other industry advocates opposed AEP’s stricter rules as “discriminatory.”
Meeting the demand
There is also a risk that the reliability of the power grid will decrease as many large data center projects progress. For example, data center demand of 13 gigawatts, which AEP considers a more accurate figure, is equivalent to about a dozen large nuclear power plants. The infrastructure required to meet this demand in power plants and transmission lines is enormous, the utility said.
PJM’s solution is for data centers to deny grid connection requests if there isn’t enough power to provide them, said Bowring of Monitoring Analytics. Data centers can either wait until enough power is available to them, or they can bring their own generation with them and jump the line, he said.
Monitoring Analytics submitted an application Complaint to the Federal Energy Regulatory Commission Last week we called on PJM to adopt this approach.
“This will give data centers a clear incentive to bring their own products. [own] “This formula will also help remove uncertainty in demand forecasts because data centers are less likely to pay for infrastructure if they are not serious,” Bowring said.
In its complaint, the watchdog warned FERC that otherwise costs from consumers’ data center demand will continue to rise.




