FactHub reusable chart

Each of the five largest AI data centers matches 170,000 to 770,000 U.S. homes

Annual electricity of the five largest operating AI data centers (mid-2026), expressed as average U.S. homes, at full and half IT load.

Credit: FactHub Research Team; capacity data from Epoch AI (CC BY 4.0); household benchmark from U.S. EIA

Node: chart:largest-ai-data-centers-homes

Dot chart of five AI data centers. At full IT load they match about 768,000 (Colossus 2), 739,000 (New Carlisle), 516,000 (Fairwater Atlanta), 456,000 (Meta Prometheus) and 342,000 (Stargate Abilene) U.S. homes; at half load, about 384,000, 369,000, 258,000, 228,000 and 171,000. All lie above a reference line at 100,000 homes.

Hover over the chart or focus a point to inspect its values.

Every one of the five largest operating AI data centers uses as much electricity as at least 170,000 average U.S. homes even if its IT equipment averaged only half its capacity, and up to about 768,000 homes at full load. Both bounds exclude cooling overhead.

What this chart asks

How many average U.S. homes use as much electricity as each of the largest AI data centers?

Why it is shown this way

  • Points rather than stacked bars, so the full-load and half-load values are read as alternatives, not as parts that add up.
  • The y axis starts at zero so distances to the 100,000-home reference line are honest.

Chart annotations

These markers provide relevant context for interpreting the chart. Their timing or position does not, by itself, establish causation.

Typical large data center: about 100,000 homes (CRS)
Threshold: 100000

Congressional Research Service, citing the IEA, says large data centers typically use as much energy as roughly 100,000 households.

Why this is marked: Reference level from the Congressional Research Service (R48583, June 2025) so readers can compare the largest AI sites with a typical large data center; not part of the dataset.

Data and accessibility

Download chart data · Open accessible table · Download PNG · Download SVG

Dataset largest-ai-data-centers-power, analysis household-equivalents

Sources

This chart draws on the sources of the dataset versions it visualizes.

Largest operating AI data centers: power capacity and household-equivalent electricity, mid-2026

  1. Epoch AI, 'The record for power capacity in a single data center has doubled every 10 months' (Sep. 4, 2026), with its Frontier Data Centers data IT power capacity estimates and estimate dates for each facility

    Retrieved October 26, 2026. CC BY 4.0. Allowed with attribution to Epoch AI

    Reliability: Academic. Epoch AI is an independent nonprofit research group that estimates data center power from satellite imagery, permits and public documents. It does not own or operate data centers.

  2. U.S. Energy Information Administration, 'How much electricity does an American home use?' Average annual electricity sold per U.S. residential customer (10,791 kWh, 2022)

    Retrieved October 26, 2026. Public domain (U.S. government work). Unrestricted

    Reliability: Government. EIA is the statistical agency of the U.S. Department of Energy; the figure comes from utility sales reports.

Limitations

  • IT power excludes cooling, so whole-facility use is higher than shown.
  • Actual utilisation of each site is not public; full and half load bracket a plausible range rather than measure it.
  • Capacity values are Epoch AI estimates dated May to August 2026.

Articles and datasets using this chart

Version details

The exact citing versions, and the version of this chart each cites.

  • Article largest-ai-data-centers-electricity — version 1 (published) cites version 1