According to a brief released this month by the Center for Climate and Energy Solutions, U.S. electricity demand is rising for the first time in twenty years, due partly to the power requirements of AI data centers, the Center for Climate and Energy Solutions reported in a brief published this month.
The analysis, titled Beyond the Boom: Assessing the Role of AI Data Centers in Scaling Clean Firm Power Technologies, lands as hyperscalers pour billions into infrastructure. Its central finding is blunt: when operators need reliable power today, they still choose conventional firm resources — overwhelmingly unabated natural gas — because clean firm alternatives cost more upfront and take longer to build.
“Conventional firm resources remain the preferred option for meeting current AI data center operational requirements,” the brief says, defining clean firm power as low- or zero-carbon technologies that can dispatch on command — advanced nuclear, next-generation geothermal, long-duration storage paired with renewables, and natural gas with carbon capture.

AI clusters run at near-constant, high utilization; a training job that sits idle burns capital without producing value. That flat-out, round-the-clock load profile separates these facilities from typical commercial buildings or even earlier cloud data centers. Solar and wind fill only part of the need. Batteries cover short gaps. For the long stretches when variable output drops, operators lean on sources that can run on command.
Shattered notes that this demand shape is exactly why climate-focused groups watch AI electricity growth so closely. A purchaser with deep pockets, long planning horizons, and a need for 24/7 supply looks like the ideal early customer for technologies that struggle to find anyone willing to pay a premium. The brief tests whether that match holds up in practice and concludes it does not — not yet.
C2ES names four hyperscalers in the brief, though the public summary does not detail their individual procurement strategies. Shattered reports that the analysis examines what the brief confirms, what it leaves open, and how the findings fit with other power stories including the emissions footprint of U.S. hyperscale data centers.
The omission of company-level data highlights a larger lack of transparency: the ultimate climate impact of AI’s growth is still uncertain because the fuel mix behind new capacity is still being decided in real time.
Tax credit extensions and permitting reform are identified as the policy levers that could help clean firm technologies compete and scale more effectively. Without them, the cost and timeline disadvantage persists, and gas-fired generation continues to win the next megawatt.
C2ES will present the report’s findings on December 1, 2026, from 12:00–1:30 p.m. ET, hosting a fireside chat with Senator Sheldon Whitehouse and C2ES President Nat Keohane, followed by a panel with perspectives from hyperscalers, utilities, and clean firm technology providers.
- Construction Employment up 1.5% Year Over Year in October, BLS Says
- FederalLawyers.com Ranks Wyoming Business Debt Settlement Firms for 2026
- 3.6% for Social Security, 2.6% for FERS, and a Bill Still Pending for VA (daybreakwire.com)
- Honoring Dolores Huerta: The Legacy of a Civil Rights Icon and the Power of Sí Se Puede (archyde.com)