America's Power Grid Faces 'Crisis Stage' as AI Data Centers Drive Electricity Costs Up 267%

The race to dominate artificial intelligence is pushing America's electrical grid to its breaking point. PJM Interconnection, the nation's largest power grid operator serving 65 million customers across 13 states, has issued a stark warning: the system will be six gigawatts short of its reliability requirement by 2027—roughly equivalent to the electricity demand of Philadelphia.

"It's at a crisis stage right now. PJM has never been this short," said Joe Bowring, president of Monitoring Analytics, which serves as the independent market monitor for PJM.

The culprit is clear. Data centers, powered by the insatiable computational demands of artificial intelligence, are consuming electricity at rates that have caught utilities, regulators, and grid operators off guard. And American households are paying the price.

The Bill Arrives

Residential electricity prices rose 5.2% in October compared to the same period last year, and the federal Energy Information Administration forecasts another 4% increase in 2026. But those national averages mask far more dramatic increases in regions hosting major data center clusters.

Electricity costs for areas near data centers have increased by as much as 267% compared to five years ago, according to a Bloomberg News analysis. The cost burden falls disproportionately on residential customers and small businesses, who lack the negotiating leverage of large industrial users.

Data centers tied to PJM's grid added $6.5 billion to the cost of procuring power supplies in a recent capacity auction. Total costs attributable to data centers reached $23.1 billion for the three-year period from June 2025 through May 2028, representing 49% of the $47.2 billion total.

Strange Bedfellows in Opposition

The data center boom has accomplished something rare in American politics: uniting Bernie Sanders and Ron DeSantis in opposition. The Vermont senator has called for a national moratorium on data center construction, while the Florida governor has raised concerns about the strain on state infrastructure.

The alignment signals a political reckoning brewing over the AI industry's impact on electricity prices, grid stability, and the labor market. Unlike manufacturing investments that create broad employment, data centers generate relatively few permanent jobs while consuming enormous resources.

"We all benefit from AI, but we shouldn't have to subsidize Big Tech's electricity bills. Regular families are paying more so these companies can train their algorithms."

— Consumer advocacy organization representative

The Scale of the Problem

Data centers are projected to consume 6.7% to 12% of U.S. electricity by 2028, up from 4.4% in 2023, according to a December 2024 Department of Energy report. BloombergNEF's latest forecast projects data center power demand will hit 106 gigawatts by 2035—a 36% increase from projections made just seven months earlier.

The surge reflects not just more data centers, but larger ones. New AI-focused facilities require substantially more power than traditional cloud computing centers. A single advanced AI training cluster can consume as much electricity as a small town.

Corporate investments in artificial intelligence, especially in data center construction, drove over one-third of U.S. GDP growth in the first nine months of 2025. But that economic contribution comes with mounting infrastructure costs that taxpayers and ratepayers must bear.

The Path Forward

The Trump administration and a consortium of northeastern governors have asked PJM to hold an emergency power auction that would require tech giants to pay for the surging costs their data centers create. The proposal would shift some burden from residential customers to commercial operators.

Industry experts predict that in 2026, data centers will need to play a more active role in stabilizing the grid through load flexibility—reducing power consumption during peak demand periods. Some companies are exploring on-site generation through natural gas or nuclear microreactors.

But solutions remain years away, while the problem is immediate. Grid operators face the challenge of maintaining reliability with aging infrastructure while integrating unprecedented new demand. For American households, the result is simple: higher electricity bills with no end in sight.

The irony is inescapable. The companies promising to make our lives more efficient through artificial intelligence are simultaneously making a fundamental necessity—electricity—more expensive and less reliable.