AI’s relentless energy demands will strain the power grid for the next decade, regulatory group says

A report from an industry watchdog warned that North America’s electric grid will be under strain for the next decade as demand from artificial intelligence threatens its stability.

With consumption on the rise while many power plants in the US and Canada face closures, North America’s grid could encounter “mounting resource adequacy challenges,” the North American Electric Reliability Corporation said.

“The trends point to critical reliability challenges facing the industry: satisfying escalating energy growth, managing generator retirements, and accelerating resource and transmission development,” the group said in its 2024 Long Term Reliability Assessment report.

The non-profit watchdog is subject to oversight from the Federal Energy Regulatory Commission and regularly assesses the grid’s security.

In its latest outlook, NERC said that power demand growth is at its highest point in two decades, as crypto- and AI-driven data centers connect rapidly to the grid. Electric vehicles and the rising use of heat pumps are further amplifying demand.

NERC expects summer demand to peak by 15%, climbing 132 gigawatts. The forecast marks a steep increase from last year’s expectations of an 80 gigawatt rise. Peak winter demand will jump 18%.

Meanwhile, US and Canadian regions are increasingly at risk of shortfalls in power supply, as key sources of electric generation are taken offline or are simply inadequate to avoid load-loss events.

“Accelerated retirements of the existing coal, natural gas, and nuclear generators can have a profound and negative effect on the resource adequacy and reliability of the [bulk power system] in the next 10 years,” NERC said. “The lack of dispatchable resources and diverse generator fuel types in the interconnection processes makes the future resource mix look alarmingly unreliable.”

The massive power demands from AI are well-known, which explains why investment has rushed into infrastructure and utility projects. Even nuclear power companies have garnered market attention amid an AI-led rush to revitalize the sector.

Recently, Microsoft CEO Satya Nadella acknowledged that power constraints have become the top challenge of the AI buildout. The tech titan is among the companies that have turned toward nuclear power.

In a July note, Bank of America cautioned that the tech sector’s power consumption would trigger a rebound in utility prices. Though costs shrunk earlier this year, AI’s massive electric demand will likely drive utility payments higher over a multi-decade timeframe.

By 2026, the bank said it anticipates an additional 18 to 28 gigawatts of electric capacity will be required.