For two decades, American electricity demand sat still. That era is over, and Meketa Investment Group1 authors Gerald Chew, Adam Toczylowski and Lily Hillis frame the consequence as a problem of timing, not scarcity. "The core challenge is one of timing," the paper states, and the authors identify "a widening gap between accelerating electricity demand and a supply system that takes years to expand."
Why Is the Meter Moving Again?
Demand grew 2.1% per year over the five years through 2025, after being nearly flat from 2008 to 2023. The EIA projects cumulative growth of 25%-50% by 2050, implying 50%-90% more generating capacity. Data centers lead, with EPRI scenarios spanning roughly 380 TWh to nearly 800 TWh by 2030 (see Figure 2). The authors stress the spread: EPRI's high case is "roughly twice its low scenario."
Data centers are not alone. Electric vehicles climb from about 30 TWh in 2025 to 90-140 TWh in 2030, chip fabs add concentrated load, and the South, home to major data center markets, grew 6.0% since 2020 against 3.1% nationally. Efficiency long masked this: from 2000 to 2025 GDP rose more than 70%, yet electricity rose about 17%. Now, the authors write, "the new sources of increasing load are now layered on top of this baseline, compounding the effects over time."
Who Supplies the Power?
"No single source has carried the US power system, and long-term projections suggest none will for the foreseeable future." The EIA baseline has gas near 40% of 2050 generation, renewables above 45% and nuclear about 13%. Gas is dispatchable, which matters because the grid, the paper notes, "must be built for the busiest hour, not the average hour."
Renewables are the fastest-growing source, yet solar contracted to match a data center's annual use "typically covers only 35% to 45% of that facility's demand on an hourly basis." They grow, the authors write, "alongside other sources rather than in place of them."
Nuclear shows the sharpest divergence. One estimate puts restarts near $3,100 per kilowatt against $6,400 to $12,700 for new builds, yet only a few reactors qualify. The 400 GW federal goal needs about 15 GW a year from 2030 to 2050, against a historical peak of 10.5 GW. Small modular reactors promise more, but their advantages "have not yet been demonstrated on a commercial scale."
Where Does the Power Get Stuck?
About 2,060 GW of projects wait to connect, more than one and a half times the 1,280 GW installed, and only about 19% of projects queued from 2000 to 2020 reached operation. "In many regions, the binding constraint on new supply is not the availability of energy resources but the grid's ability to deliver power where and when it is needed." Transformer lead times reach about 128 weeks, gas turbines may arrive in 2028 or later, and over half of energy employers report serious hiring bottlenecks. Minerals follow that logic: "The main constraint is often not geology."
The 1973-1974 embargo exposed import dependence, shale reversed it, and the US became a net energy exporter in 2019. To be clear, the lesson has widened: "Domestic resource abundance offers limited protection if essential processing or components are concentrated elsewhere." Policy compounds the risk: the 2025 law ends EV credits and phases out wind and solar support, and the authors warn: "Incentives can change faster than the assets they support can be built."
What Is Already in the Price?
"Higher demand alone does not make a good investment." The paper points to assets that relieve bottlenecks, including grid equipment, transmission, midstream and operating plants with interconnection, because "capital alone cannot create grid capacity." To be fair, the case has cracks. "Utility load forecasts may also be inflated," retail prices rose 5.3% in 2025, and "Assets built for demand that does not arrive could become stranded." Valuations may already reflect the growth, and the authors add: "Entry price will matter as much as the theme itself."
The conclusion is plain: "How quickly the US clears these bottlenecks will set the pace of the buildout."
Five Takeaways for Advisors and Investors
- Respect the range. Demand growth is real, but EPRI's 2030 scenarios differ by roughly double.
- Own deliverability. Grid equipment, transmission, midstream and mineral processing relieve the constraint.
- Diversify the supply stack. Gas, nuclear, renewables and storage are complements, not substitutes.
- Treat policy as risk. Favour long-term contracts with investment-grade counterparties.
- Price and selection decide. Scrutinize entry valuation, contract structure and manager skill.
Footnote:
1 Chew, Gerald, Adam Toczylowski, and Lily Hillis. "Powering America's Future." Meketa Investment Group, Nov. 2026, https://meketa.com/leadership/powering-americas-future/. Accessed 10 Nov. 2026.
