Regional power generation and related infrastructure have been under pressure for years, with causes varying region to region, highlighted by prominent disruptions or threats thereof. Indeed, as depicted below power outages in the United States have increased both in frequency and duration over the last decade. US Energy Information Agency data tracking both metrics noted meaningful increases from 2013 through 2021.1
Regionally, there have been notable, high profile and costly grid failures driven by severe weather events. In the Electricity Reliability Council of Texas (ERCOT) market, these challenges were on display during Winter Storm Uri in February 2021 which saw ~48.6 percent of generation forced out at one point due to extreme weather conditions.2 Challenges with natural gas availability due to freezing pipes and inadequately weatherized power assets (thermal and renewable) both played a role, resulting in over 4.5 million Texans losing power, some for several days, and at least 210 deaths. The events of Winter Storm Uri mirrored those from 7 years earlier in January 2014, when the polar vortex wrought havoc on power grids—particularly the Pennsylvania New Jersey Maryland (PJM) regional grid.3
Where extreme cold challenged the PJM and ERCOT grids, it was extreme heat that brought the California Independent System Operator (CAISO) grid to its knees in August 2020. Those blackouts represented the first rolling blackouts in CAISO “since 2001, when Enron and other energy traders manipulated California’s market.”4 Generation outages due to extreme heat and transmission constraints were both cited as factors.
The threat of outages looms over other regional markets. In November 2023, the North American Electric Reliability Corp (NERC) warned that extreme cold weather could threaten 180 million Americans and Canadians due to lacking natural gas infrastructure.5 Indeed, transmission constraints and limited natural gas pipeline infrastructure loom particularly large in the Independent System Operator New England
(ISO NE), where attempts to add natural gas infrastructure (like the cancelled Northeast Energy Direct project)6 and transmission (like the similarly cancelled Twin States Clean Energy Link)7 have often been unsuccessful resulting in system constraints.
The aforementioned challenges for regional grids occurred despite it being a time when demand growth—and the resultant incremental strain on those systems—had been comparatively muted. Enter the onset of potentially significant growth. Concerns about this trend began several years ago, with the NERC noting in its 2022 Long-Term Reliability Assessment highlighting that “projected growth rates of electricity peak demand and energy in North America are increasing for the first time in recent years.”8 Electrification and data center growth, primarily from crypto operations, were noted at the time. Inside of two years later, the expectations for growth have accelerated. A report by Grid Strategies filed with the Federal Energy Regulatory Commission (FERC) entitled “The Era of Flat Power Demand is Over” pointed to projected demand growth accelerating to a rate of 4.7 percent from a prior estimate of 2.6 percent year-over-year for expected peak demand in 2028, primarily as a result of data center demand broadly and AI specifically.9,10 Markets like PJM were already anticipating significant growth in peak load requirements from electrification before contemplating significant additional demand pull from the addition of data centers as depicted below.
More worryingly, Grid Strategies noted that such forecasts are “likely an underestimate: several more recent updates are adding additional GWs to that forecast. Next year’s forecast is likely to show an even higher nationwide growth rate.”11 Goldman Sachs, in a recent report on the implications of data center and AI-related demand, drew a similar conclusion noting “an acceleration in U.S. electricity demand CAGR to 2.4 percent through the end of the decade from 0% in the last decade.”12
With regional grids straining following a period of relatively muted demand growth, it is fair to wonder how they will perform absent significant additional investment. Indeed, Goldman Sachs estimates $50 billion of additional investment in U.S. power generation capacity will be needed to meet data center demand growth.13 Indeed, there are a number of areas that are expected to see increased investment as a result of this forecast growth.
The U.S. power market is approaching a period of potentially significant growth, well above what it experienced thus far this century. With grid instability already a challenge in many regions, the need for additional investment in a range of solutions is coming into focus. Power generation, energy infrastructure and energy storage all are likely to see increased investment as a result. Service providers and efficiency solutions are also likely to play an important role in the coming decade as the United States seeks to embrace—and manage— a new wave of power demand.