{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigzmkp4xjkxjfegnaequdktcddrzx6wmwseunzweo7qgj7xuhb6fq",
    "uri": "at://did:plc:mg5ozsljpp6t5b4lvwys4t72/app.bsky.feed.post/3mlqjdmrznbt2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreicmsn5mpkl6u6urdvmeej3ldz4iyyxzur7ubapqoghfjh5wdlyaau"
    },
    "mimeType": "image/png",
    "size": 442643
  },
  "description": "Utilities are struggling to connect large data centers quickly enough to maintain reliability, panelists said.\n",
  "path": "/ai-power-demand-is-outpacing-grid-buildout/",
  "publishedAt": "2026-05-13T14:30:31.000Z",
  "site": "https://broadbandbreakfast.com",
  "tags": [
    "25 percent increase in U.S. electricity production"
  ],
  "textContent": "WASHINGTON, May 13, 2026 – Artificial Intelligence and large-scale data centers are driving electricity demand faster than utilities can expand generation and transmission infrastructure, energy regulators said Tuesday.\n\nPanelists speaking at a United States Energy Association event warned that gigawatt-scale data centers are increasingly requesting grid interconnections within one to three years, while transmission projects can take seven years or more to complete.\n\n“The speed and the size of these loads” are creating major reliability challenges, said **Fritz Herve,** vice president of government affairs at the North American Electric Reliability Corporation**.**\n\nAccording to Herve, summer peak electricity demand is expected to increase by 224 gigawatts over the next decade, driven largely by new data centers and AI infrastructure\n\nIndustry officials warned last year that AI infrastructure could require a 25 percent increase in U.S. electricity production over five years as utilities race to connect new data center loads\n\nHe also warned that some large data center loads can rapidly turn on and off, creating frequency fluctuations that threaten grid stability.\n\n“When you’re talking about loads that are in the 1,500 megawatt size or bigger, those can cause frequency variances on the system,” Herve said.\n\nThe reliability organization recently issued new guidance for large-load planning and operations and is proposing, for the first time, to formally register large-load customers under its regulatory framework.\n\n**Tony Clark,** executive director of the National Association of Regulatory Utility Commissioners**,** said states were increasingly being forced to manage the infrastructure and cost implications of rapid data center growth on a region-by-region basis.\n\n“There’s no one way of doing things all the way across the country,” Clark said.\n\nClark said states were experimenting with new cost-allocation mechanisms aimed at preventing existing utility customers from absorbing the risks associated with large new electricity loads.\n\nThose mechanisms include minimum contract terms, collateral requirements, demand charges, and “take-or-pay” agreements requiring large-load customers to share infrastructure risks.\n\nClark also warned states were closely watching whether federal regulators could eventually override state authority over retail electricity load and infrastructure cost allocation.\n\n“There is some concern,” Clark said, referring to possible federal action involving large-load interconnections.",
  "title": "AI Power Demand Is Outpacing Grid Buildout",
  "updatedAt": "2026-05-21T21:47:44.886Z"
}