Power Demand Forecasts Revised Up for Third Year Running, Led by Data Centers
Grid Strategies LLCJohn D. Wilson, Sophie Meyer, Zach Zimmerman, Rob Gramlich01 Nov 2025all-rights-reservedRead the original
No copyright notice or reuse statement found in the document. Grid Strategies is a commercial consultancy; all rights reserved assumed from publisher type. Widely cited by convention, but no licence has been granted.
Grid Strategies' third annual review of US utility load forecasts finds projected demand growth revised upward again. The five-year forecast of utility peak load growth has risen more than sixfold over three years, from 24 GW to 166 GW. Electricity consumption is forecast to grow at an average 5.7 percent a year over five years, a cumulative increase of around 32 percent by 2030, while peak demand grows at a slower 3.7 percent annual rate. Data centres are identified as the dominant driver, accounting for roughly 55 percent of demand growth in utility load forecasts. The authors add a significant caveat about double counting and speculative projects: comparing utility filings against market analysts' published assessments, they estimate the data centre share of utility forecasts is overstated by roughly 25 GW. Growth is geographically concentrated, with six regions including CAISO, ERCOT, PJM and SPP representing over 80 percent of projected growth.
Contested
Quotation
“"Over the past three years, the 5-year forecast of utility peak load growth has increased by more than a factor of six, from 24 GW to 166 GW."”
Grid Strategies LLC · all-rights-reserved · Read the original
All rights reserved. Only this short quotation is reproduced here.
The MyGreenSuit briefing
Commentary by MyGreenSuit, not the words of Grid Strategies LLC.
Sixfold in three years. That is either a forecasting profession that was badly wrong in 2022 or one that is badly wrong now, and both possibilities should make a planner uncomfortable.
The most honest thing in the report is the authors' own caveat, which almost nobody quotes. Comparing utility filings against independent market analysis, they estimate the data centre component is overstated by around 25 gigawatts through double counting. A developer shopping one project across four utilities appears four times.
So the headline number and the footnote point in opposite directions, and the report says so. That makes it a more useful document than one that does not.
What is not in dispute is concentration: six regions account for more than 80 percent of projected growth. This is not a national load problem. It is a CAISO, ERCOT, PJM and SPP problem.
One caution before quoting it as neutral — the sponsor of this edition is not identified in the copy I have. Worth establishing.