Why AI is arriving at the most difficult moment for North America’s grid
Get the Energy newsletter
Daily energy & climate — solar, EVs, oil, the policy fights and tech bets shaping the transition. Free.
- NERC projected 224 GW of summer peak demand growth over the next decade in its 2025 Long-Term Reliability Assessment — the highest 10-year peak demand growth since 1995 — identifying data centers and large-load customers as key drivers.
- Lawrence Berkeley National Laboratory projected U.S. data center electricity consumption will grow from approximately 176 TWh in 2023 to between 325 and 580 TWh by 2028, but flagged uncertainty over timing, location, and load concentration as the harder planning question.
- Interconnection queues are shifting from a technical bottleneck to an economic one, with grid readiness now influencing hyperscale data center and advanced manufacturing siting decisions as heavily as land, workforce, or capital availability.
- Utilities are abandoning single-forecast planning in favor of evaluating multiple plausible demand scenarios simultaneously, using advanced analytics, digital workflows, high-performance computing, and AI-enabled planning tools alongside phased interconnection, demand response, and flexible service agreements.
- Siemens, NVIDIA, and Dominion Energy are co-hosting a discussion titled "Planning the Grid at the Speed of AI: From Interconnection Queues to Grid Readiness" focused on how leading organizations are responding to large-load growth.
Why it matters: The 224 GW demand jump is the easy part — the hard part is no one knows where or when those loads will land, so utilities that clear interconnection queues faster and offer flexible service agreements will capture hyperscale capital, while regions still running on traditional forecasting timelines risk watching major data center investments go to faster-moving competitors.
Ask SkimNews
