Utilities around the country expect peak demand to grow by 128 GW, or 15.8%, to 947 GW by 2029, according to a report Grid Strategies released Dec. 5.
The new figure represents an 11% increase from the organization’s previous five-year prediction, made about a year ago in a similar report. (See Grid Planners Predict Sharp Increase in Load Growth.)
“Power demand had doubled last year from the prior year; lo and behold, it has doubled again,” Grid Strategies President Rob Gramlich said on a Zoom call with reporters. “So, it’s a lot. It’s hard to think of an electricity industry process that is not affected by power demand, whether it’s utility integrated resource plans, rate cases, transmission planning, resource adequacy, you name it.”
Six regions are the primary drivers of load growth, including ERCOT with 43 GW, PJM with 30 GW, and the state of Georgia with 13 GW. But Grid Strategies Vice President and report co-author John Wilson noted that the growth within those regions is really centered in Dallas, Northern Virginia and Atlanta, respectively, as data centers and new manufacturing are locating there.
“Putting this in broader historical context … assuming this growth forecast is correct, we are now looking at the latter half of this decade showing 3% annual average load growth,” Wilson said.
The last time the industry saw load growing at that clip was the 1980s, and given the decades since then, 3% is larger overall, Wilson said. It will take more overall infrastructure investments to meet that level of growth now, he added.
Load forecasts were also very high for the internet boom in the 1990s, but efficiency won out, and the largest forecasts did not pan out, Gramlich said.
“I think the overall message I got in talking to experts on this topic is that efficiency is not what’s going to drive this,” Wilson said. “To the extent that energy efficiency is improved, it may just lead to more computing demand and not reduced energy demand. And I think that’s a pretty strong message we got from the folks who are more expert on this topic than we are.”
The 128-GW figure includes 67 GW from FERC Form 714 filings on load growth, which is up from 39 GW last year and 23 GW in 2022. Grid Strategies calls this the “official nationwide forecast.” Even by itself, it represents an 8.2% increase over five years.
The remaining 61 GW is based on “recent updates” from ERCOT, PJM and Georgia Power, the report says. Wilson said Grid Strategies did not have access to that quality of data from all around the country, but the additions from those three regions were so large that it warranted inclusion in the report.
Forecasts for data centers vary, with the industry expecting it will grow by 65 GW, while updated utility forecasts suggest it could hit over 90 GW, the report says.
The numbers are just forecasts, Wilson noted, and given that they are based on the addition of large loads from a few customers, utilities could invest to meet that higher demand only to see data centers shut down after a few years of operating, leaving other customers to foot the bill for now unneeded upgrades.
“That’s why this issue has become such a hot topic in a lot of jurisdictions,” Wilson said. “Another big hot topic that you may not be familiar with is the impact of data centers on reliability, and this is something that NERC is paying very serious attention to.”
Data centers can be very sensitive to fluctuations in voltage and other key grid power quality metrics, which means if there is a fault on a line — such as if it is struck by lightning — then their backup systems might kick in and move data centers off the grid.
“If you’ve got an 80-MW data center, and it’s out there all alone in a region on the grid, that’s probably not a huge problem for the grid,” Wilson said. “But if you’ve got several 400-MW data centers, and they’re located in proximity to each other, and they all trip off the grid at the same time, now you’re talking about the equivalent of a 1-GW-plus power plant needing to essentially disappear from the grid instantly, and that’s a real challenge for the grid.”
Manufacturing forecasts are unavailable, but indicators suggest its growth could add 20 GW of demand, and electrification could add another 20 GW. Electrification would have a bigger impact if the report went into the 2030s, by which point states like New York expect to be winter peaking as more buildings are heated by grid power, Wilson said.
“When you get to 2035, the winter peak becomes so large that now the system becomes winter peaking, and the building electrification now is the big driver of the peak,” Wilson said. “So, this is why electrification is in a lot of regions something that is a phenomenon of the 2030s and not of the 2020s; it’s just not driving the peaks.”