ERCOT Projects Texas Peak Demand Could Quadruple to 367 GW by 2032 as Data Centers Drive Over 60% of Growth, but Grid Operator Warns Forecast Is Likely Inflated
ERCOT's preliminary 2026-2032 load forecast filed April 15 projects peak demand reaching 367,790 MW, more than four times the current record, with data centers accounting for over 60% of the projected growth. CEO Pablo Vegas and the PUCT say the number is almost certainly too high.
Overview
The Electric Reliability Council of Texas (ERCOT) filed a preliminary long-term load forecast on April 15, 2026, projecting that peak electricity demand in its footprint could climb from the current all-time record of 85,508 megawatts to 367,790 megawatts by 2032, according to Utility Dive. That is a more than four-fold increase in six years, driven overwhelmingly by the hyperscale data-center buildout concentrated in Texas.
The filing arrived with an unusual caveat: ERCOT itself, and the Public Utility Commission of Texas (PUCT), told regulators the number is almost certainly too high. The forecast is based on interconnection requests from transmission providers, many of which are speculative, and the agency has been directed to file a revised version.
What We Know
ERCOT’s preliminary base forecast projects a relatively modest rise in organic peak demand from about 98,087 MW in 2026 to roughly 111,318 MW by 2032, but jumps to the 367,790 MW headline figure once large and medium loads submitted by transmission companies are folded in, Utility Dive reports. The current ERCOT record of 85,508 MW was set in August 2023.
Data centers are the dominant driver. Bloomberg, whose wire account of the filing was picked up across the trade press, reported that data centers account for more than 60% of the projected increase and that accommodating the forecast would require the equivalent of almost 300 new nuclear reactors.
The sheer size of the interconnection pipeline is part of why the forecast looks so extreme. ERCOT’s large-load queue jumped nearly 300% in 2025 and totaled about 233 GW of capacity under review, with more than 70% of those requests coming from data centers, Utility Dive reported. The existing review process was designed to handle 40 to 50 large loads at a time; ERCOT received 225 new requests by mid-November 2025 alone.
ERCOT’s Own Caution
ERCOT President and CEO Pablo Vegas told regulators the grid operator does not believe the preliminary forecast reflects what will actually get built. “Texas is experiencing exceptional growth and development, which is reshaping how large load demand is identified, verified, and incorporated into long-term planning. As a result of a changing landscape, we believe this forecast to be higher than expected future load growth,” Vegas said in statements accompanying the filing, per Utility Dive.
Chad Seely, ERCOT’s senior vice president of regulatory policy, said the organization “has concerns with using the preliminary load forecast values” for reliability assessments or resource-adequacy planning, according to Utility Dive. PUCT officials have agreed the projection is inflated and have asked ERCOT to submit a revised forecast.
Industry voices echoed that skepticism. Travis Kavulla of Base Power told Bloomberg that “that is not the actual forecast because it can’t” realistically materialize, arguing only the data-center developers themselves know which of their projects are real.
Why the Numbers Are Hard to Pin Down
Texas has become the single largest US market for data-center interconnection requests, and the PUCT is in the middle of implementing new rules under Senate Bill 6 to vet them more carefully — an effort triggered by the same queue-inflation problem on display in the April 15 forecast, Utility Dive has reported. The disconnect between announced projects and delivered gigawatts is already visible elsewhere in the industry. The Machine Herald previously reported that Sightline Climate estimates 30-50% of US data centers slated for 2026 will be delayed or canceled because of transformer and electrical-equipment shortages, which suggests Texas’s queue will shrink considerably even without regulatory filtering.
What We Don’t Know
- Which of the 233 GW of queued large loads will actually sign firm power contracts and begin construction. Many requests in the ERCOT queue are duplicate submissions from developers hedging across multiple utilities.
- How much of the data-center load in the queue survives once SB 6’s financial-security and disclosure rules take effect, and whether that filter will reduce the forecast by a factor of two, three, or more.
- When ERCOT will file the revised forecast PUCT has requested, and how much lower the revised headline number will be.
- Whether the PUCT’s new interconnection process can be finalized before the 2027 Texas legislative session, as officials have targeted, or whether continued queue growth forces a mid-cycle rewrite.
Analysis
The ERCOT filing is notable less for the 367,790 MW headline than for the public self-repudiation attached to it. Grid operators generally prefer to err on the side of over-forecasting because under-building capacity carries reliability risk; having the CEO, general counsel’s office, and the state regulator simultaneously describe the filing as too high is a sign of how badly the traditional interconnection-queue-as-forecast methodology has broken down under hyperscale AI demand.
It also underscores a structural problem the industry now faces across every major US grid: no one knows what fraction of announced data-center projects is real. Texas’s 233 GW large-load queue, reported by Utility Dive, is widely understood to include significant double-counting from developers shopping the same project across multiple utilities. Until SB 6’s financial-security rules — and analogous filters elsewhere — actually bite, headline load forecasts will keep producing numbers that are simultaneously the official view and, in ERCOT’s own words, higher than expected future growth.