“We need to move ... from the broad concept of energy efficiency to, more specifically, demand efficiency. Are we using energy, and saving energy, during the times of day, and in the locations in which it matters?” McGee Young, founder and CEO of San Francisco-based climate tech software company WattCarbon, said during a panel discussion Wednesday.
The conversation was hosted by the American Council for an Energy-Efficient Economy (ACEEE) to consider some of the findings in its new report, Faster and Cheaper: Demand-Side Approaches for Rapid Load Growth.
The study, and the urgency around addressing strains on the grid which are only expected to increase, is driven — most immediately — by the growth of data centers to provide the computing muscle for AI technologies. Those data centers can use expansive amounts of electricity and water, setting off alarm bells among utility regulators, rate payers, electric equipment suppliers and the myriad stakeholders around environmental sustainability.
“We expect more of this load growth from the data centers to continue. And we need to figure out how to serve that load,” Katherine Peretick, a member of the Michigan Public Service Commission, said during the panel. Peretick also chairs the Electric Power Research Institute Advisory Council. “We’re starting to see these electric supply constraints in some parts of the country. And we’re seeing bottlenecks in the traditional ways of solving these constraints.”
In his discussion of demand-side efficiencies, Young used a pipe metaphor, with the fluid flowing through the pipe analogous to electricity flowing through the grid. If peak demand — the amount of water needed — keeps increasing, “we’re going to need a bigger pipe,” Young said. “And that bigger pipe is going to cost everyone more money.”
“What if instead of building a bigger pipe, we just reduced the peak flow,” he said. “And we served more of those electrons during other times of the day when the pipe wasn’t quite so full.”
Energy efficiency is not the same as demand efficiency, Young pointed out, noting that creating energy efficiency improvements was the right paradigm when generation costs were the main drivers of electric power costs. The thinking was: reduce the amount of power an appliance uses, and the utility will have to generate less electricity. However, the main drivers of electric utility costs today are on the transmission and delivery side of the equations — the pipes — according to Young and the data in the ACEEE report.
Energy efficiency was viewed as any time of day the savings happened, Young said; “it all kind of counts the same.” Demand efficiency measures, he said, are aimed at more targeted savings, and shaving off demand during peak periods.
Electric load growth has been mostly flat for the last 20 years, owing largely to improved efficiencies, said Mike Specian, ACEEE utilities manager and one of the authors of the report. Specian was formerly with the U.S. Department of Energy, where he facilitated behind-the-meter innovations to make buildings more efficient and resilient.
“But that’s changing,” he told the panel, as he listed the growth of data centers, the electrification of transportation, and the re-shoring and electrification of industry.
By mid-century ACEEE anticipates 2,000 terawatt hours of new electricity growth, which is roughly half the current electricity consumption in the United States today.
In the next five years, Specian said, data centers “will be the overwhelming driver of new load growth.” In the next decade the need for electricity is expected to increase 20 percent to 50 percent, according ACEEE, with peak demand increasing 19 percent to 35 percent.
Increased efficiencies combined with load flexibility measures could help to absorb some of the increases in electricity growth, Specian said.
“There is a reservoir of demand savings, ready to be tapped. It is less expensive to do so than supply-side alternatives. It’s faster to deploy, and it better protects rate payers and the planet,” Specian said.
Some regions have already moved in the direction of “smart grids,” where data informs new ways to manage the electric load. EPB, a municipal utility serving a 600-square-mile area including Chattanooga, Tenn., uses its fiber broadband infrastructure to enable new “smart grid” technologies around automation and other features to improve efficiencies and reduce outages.
Smart meters can produce the minute-by-minute data to structure demand-side efficiency programs and strategies, Young said.
“It’s far past time that we start to do hourly measurements and verifications to show when exactly during the day these savings are showing up,” he said. “If we’re going to think about demand-side energy resources as part of this broader grid mix, we have to scale this.”
