Utilities face significant forecasting risks from large loads, prompting the industry to develop strategies to eliminate counting of speculative projects, experts said during a Western Interstate Energy Board webinar.
Natalie Frick, deputy department leader of energy markets and policy at the Lawrence Berkeley National Laboratory, and Shana Ramirez, director at Energy and Environmental Economics, discussed the challenges of large loads during a Sept. 12 webinar.
Citing data from the Electric Power Research Institute, Frick noted that all 25 utilities participating in a study on load forecasting reported challenges incorporating data centers into load forecasts.
About half of the utilities in the study included the full requested data center capacity in their load forecast, a third included a derated capacity and six utilities did not include any of the data center requests in their forecast at all, Frick said.
“All of the utilities that they interviewed identified that they’re facing challenges with incorporating data centers into load forecasts, in particular because of the speculative nature of the service requests that they’re receiving,” Frick added.
There is no agreed-upon approach for load forecasting, but speculative load interconnection requests can skew the numbers across the board, Frick’s presentation slides showed.
Georgia Power forecast a 107% compound annual growth rate for commercial large load summer peak load in the 2025 interconnection request process. But several projects have pulled out, and net load reductions are concentrated across data center projects, “particularly those that were in the earlier stages of advancing through the interconnection process,” Frick said.
Dominion Energy forecast $1.5 billion in data center capital spending between 2025 and 2027. After staff further reviewed the interconnection requests, they removed $853 million in data center expenditures “because they identified that those were speculative,” Frick said.
“They didn’t think that … those customers had enough skin in the game to really include them in the forecast,” she added.
Also, there is no “standard large load tariff,” Frick said.
“Regulators can design the tariffs to meet their state energy goals, and those could be a variety of things,” she noted. “They could be seeking to improve or strengthen resource adequacy, affordability, attracting large load customers and also air pollutant emissions reductions. And so those are all kind of some of the motivations behind these tariffs.”
Ramirez highlighted that large load growth exposes utilities to nonpayment risks, as well as potential stranded assets and credit challenges, underlining the need for effective risk mitigation strategies.
Ramirez said there is a spectrum of what utilities are doing to mitigate those risks, ranging from “very lenient” to “very strict.”
Dominion recently proposed collateral of $1.5 million per MW for large load customers that would be included in their rate costs.
This is “much higher than what we’re seeing in other utility jurisdictions that probably have more risk, like Evergy Kansas, Evergy Missouri, where they’re just asking for two years of minimum bills as collateral,” Ramirez said.
Ramirez said there are some “best practices” utilities can implement. The practices should enable cost recovery, support responsible growth and promote fair treatment of all customers, according to the presentation.
“They should be flexible, transparent, consistent, scalable, adaptable and standardized, and [they] should align the financial security requirement with evolving risks,” Ramirez said.
