{
  "id": 20330665,
  "title": "Hyperscale data centers use 2% of US electricity",
  "url": "https://urgent.news/2026/11/16/hyperscale-data-centers-use-2-of-us-electricity-20330665",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-11-16T20:52:10.000Z",
  "source": {
    "name": "Ars Technica",
    "slug": "ars-technica",
    "url": "https://arstechnica.com/science/2026/11/hyperscale-data-centers-use-2-of-us-electricity/"
  },
  "original_language": "en",
  "account": "In recent years, the tech industry's surge in Large Language Models has driven a race to build hyperscale data centers. This rapid expansion has left many questions unanswered regarding the clean and minimal impact these facilities could have on ratepayers and the environment. Collecting accurate data on the construction of these facilities has been challenging due to the lack of comprehensive information.\n\nA new study has analyzed the energy use and estimated emissions of hyperscalers operating in the United States from May 2024 to April 2025. Although the researchers were unable to distinguish between AI compute and traditional data services, their findings offer a valuable benchmark for tracking future developments in the sector.\n\nTo gather data on these centers, the research team began with a list of 675 facilities sourced from a commercial database, company statements, and satellite imagery. They excluded any facilities that lacked a reported electrical capacity or could not be verified as large buildings equipped with essential cooling and electrical infrastructure. This process narrowed the list to 403 data centers, each boasting an average electrical capacity of approximately 40 megawatts.",
  "summary": "Study attempts to tally hyperscalers’ energy use and emissions.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}