Power lines connect to an AES Ohio substation on Tipp Cowlesville Road in Tipp City. Bryant Billing / Staff
The Ohio Manufacturers’ Association raised concerns this week about the accuracy of data center energy demand forecasts and the negative effect an overestimation of those demands could have on the market.
“Driving up customer costs in the form of increased generation and transmission is a real problem right now,” OMA President Ryan Augsburger told a Columbus ballroom full of the state’s manufacturers on Aug. 27. “And our contention is that until meaningful transparency and accountability is applied to forecasting processes, customers will see no relief.”
Much of the OMA’s annual Energy Conference conference touched on forecasting, while concerns abound about how Ohio’s energy regulators will contend with a booming data center market — chock-full of confirmed and prospective projects alike — that will impact how much electricity needs to produced and transmitted across the state.
But how much demand those data centers will bring, exactly, is a pertinent question. Forecasting in Ohio is done through the utility companies themselves; they hand over estimations to the Public Utilities Commission of Ohio, and PUCO hands them over to PJM Interconnection — the regional transmission organization responsible for flowing electricity from generators to utility companies in a 13-state area.
Those forecasts essentially tell PJM what to prepare for, with forecasts showing a huge increase in energy demand indicating that there’s a need for new transmission lines or even entirely new power generators.
The problem, as OMA’s keynote speaker and energy researcher Jonathan G. Koomey put it, is that, “Nobody has any idea, a few years from now, what energy use is going to be for data centers. They don’t.”
This is for a plethora of reasons, Koomey said. AI can become more efficient; the types of AI products on offer can change; data centers don’t run at full capacity all the time. There’s also no way to conclusively predict how many data center projects will come to fruition at any given time.
With uncertain forecasting, regulators are left with a conundrum, Koomey said.
If they work on the assumption that the demand will be extremely high and invest in improvements on the grid, but if data centers don’t end up coming online, or needing as much energy, then everyone else in the market will end up footing a higher portion of the bill, driving up costs on residents and manufacturers.
On the other hand, not being prepared for a high demand would leave Ohio’s grid vulnerable and unreliable if demand from data centers actually did spike. Koomey called this dynamic an “asymmetric risk,” where “the risk of outages is much worse than the risk of over forecasting.”
In a press release following Koomey’s presentation, OMA and Augsburger called for greater transparency in the utility companies’ forecasts, and for regulators like PUCO and PJM to stop “treating them like facts.”
“If utilities and grid operators want to spend billions chasing projected demand, then prove the demand first,” Augsburger said. “Regulators should be protecting customers, not rubber-stamping speculative spending and leaving customers holding the bag when the forecasts fall apart.”