The 2025 Long-Term Reliability Assessment (LTRA) is the annual ten-year grid-adequacy outlook published by the North American Electric Reliability Corporation (NERC). Published in January 2026 and covering 2025 data, it projects summer peak demand rising by 224 GW over ten years, 69% higher than the corresponding projection in LTRA 2024, and attributes most of that increase to new data centers for artificial intelligence and the digital economy.
Publisher: North American Electric Reliability Corporation (NERC) Published: January 2026 (assessment covers 2025 data)
Demand projections
The assessment's ten-year demand projections, and their change relative to LTRA 2024, are summarized below.
| Metric | Change vs. LTRA 2024 | 10-Year Value |
|---|---|---|
| Summer peak demand | +69% (from 132 GW) | +224 GW |
| Winter peak demand | +65% (from 149 GW) | +245 GW |
| Data-center load by 2035 | — | +18 GW (specifically AI) |
NERC characterizes the 2025 LTRA as the third consecutive annual upward revision of its demand forecast, each driven primarily by AI and data-center load. The 224 GW of incremental summer-peak demand is roughly 4–5 times total US data-center capacity circa 2024.
Attribution to data centers
NERC attributes most of the projected demand growth to data-center buildout: "New data centers for artificial intelligence and the digital economy account for most of the projected increase in North American electricity demand over the next 10 years."
Regional capacity risk
NERC identifies six regions at risk of resource shortfall over the assessment period unless planned resources come online on schedule: MISO (Midcontinent ISO), PJM, Texas RE-ERCOT, WECC-Northwest, WECC-Basin, and SERC-Central. Independent analysts highlight a 2028 shortfall risk across these regions if planned-but-not-built resources slip. NERC frames the outlook as "manageable but only if planned resources come online and on time."
The assessment has been cited as a quantification of grid-reliability strain associated with AI infrastructure, and as supporting the case for accelerating generation, transmission, and nuclear power-purchase agreements, including deals involving Constellation, Amazon–Talen, Meta–Constellation, and Microsoft–Three Mile Island.
Relationships
- supports: AI Environmental Impact, Data Center Siting / AI Power Politics, Epoch AI — How Much Power Will Frontier AI Training Demand in 2030?, RAND — AI's Power Requirements Under Exponential Growth (2025), Bain & Company — 6th Annual Global Technology Report (2025), GAO-25-107172 — Generative AI's Environmental and Human Effects.
- related: Planned Nuclear PPAs for AI, AI Bubble Debate.