Back to Research

Archive

United States

ERCOT Load Growth, Data Centres and the Texas Reliability Problem Set

Transmission towers in Clay County, Texas
Transmission towers in Clay County, Texas.Photo: Jphill19, CC BY-SA 4.0, via Wikimedia Commons

In April 2025 ERCOT published two load forecasts for Texas and told its board it would plan with the smaller one. The transmission service providers' forecast, built from the loads utilities say are coming, reaches a 2031 summer peak of 218 GW. ERCOT's adjusted forecast reaches 144,522 MW in 2031. The current record peak, set in August 2023, is about 85.5 GW.

On our arithmetic the adjusted forecast is 66 percent of the utilities' number, and it still implies a 2031 peak about 1.7 times the existing record (144,522 divided by roughly 85,500). ERCOT's report puts forecast energy growth from 2025 to 2031 at 13.6 percent a year, against 3.1 percent a year over 2014 to 2024. Electric vehicles are a small part of it: the report's EV component reaches 2,641 MW by 2031, under 2 percent of the adjusted peak.

What changed in a year

The jump comes from data centers, and from a change in what counts as a forecast load. ERCOT's board presentation shows the number of large loads backed only by an officer's letter, rather than an interconnection agreement, rising from 92 in the 2024 survey to 281 in 2025, while contracted loads rose from 64 to 132. Forecast data center growth for 2030 rose from 29,614 MW in the 2024 forecast to 77,965 MW in 2025. Texas law, through House Bill 5066, now requires officer-attested loads to be treated as reasonable in transmission planning.

ERCOT's adjustment is blunt. It delays all contract and officer-letter loads by 180 days, cuts new data center load to 49.8 percent of the amount requested, and then cuts all officer-letter load to 55.4 percent. Those discounts are based on how much requested load has historically materialized. ERCOT will use the adjusted forecast for its May 2025 capacity, demand and reserves report and for the Regional Transmission Plan, and will still accept the utilities' forecast in project reviews.

Our view

We think ERCOT is right to discount, and we think the discount should be published load by load. A 2031 peak of 145 GW is still an enormous build, and the gap between 145 GW and 218 GW is about 73 GW, not far short of the entire current ERCOT peak. Transmission planned for the lower number will be short if half the officer-letter loads arrive. Transmission planned for the higher number will be paid for by every Texas customer if they do not.

The way to narrow that gap is to make large loads carry more of the risk. Financial security posted at the interconnection request stage, and lost if the load does not materialize, would sort serious projects from speculative ones faster than any statistical adjustment.

ERCOT has limited ties to neighboring grids, so a shortfall cannot be covered by imports in the way an eastern operator might manage. Gas supply in winter storms remains the other constraint, and none of the new forecast changes that.

Sources

  • ERCOT, 2025 Long-Term Hourly Peak Demand and Energy Forecast, 8 April 2025 ercot.com
  • ERCOT, Board Item 8.1: Long-Term Load Forecast Update (2025-2031) and Methodology Changes, 7 April 2025 ercot.com
  • ERCOT, Long-term Load Forecast for 2025, Regional Planning Group presentation, 29 April 2025 ercot.com
  • Dallas Morning News, Data centers drive ERCOT's massive power demand forecast, 9 April 2025 dallasnews.com

Newsletter

Get The TEI Briefing

A weekly read on energy markets, policy and our latest research. Free, and you can unsubscribe at any time.