AILoadmapAI gigawatts, campus by campus

About AI Loadmap

An independent tracker of where the next gigawatts of AI data-center load are being built, and what stands between each campus and its first power. 146 campuses, data generated 2026-10-05.

What we do

AI Loadmap is an information service. It collects facts about announced AI data-center campuses that are already public, such as press releases, regulatory filings, permits, court records, utility announcements, satellite-based observations published by third parties, and news reporting, and arranges them by campus and by date. It then applies simple, published rules to those facts (lead times, deadlines, traffic lights, scores). Those rules are working assumptions, described on the Method page, not industry standards.

The site is operated by AI Loadmap (ailoadmap.com). It is not affiliated with, endorsed by, or acting for any company, utility, regulator or government named on it.

How the data is made

Each campus has a record of dated steps: site, permits, grid agreement, substation, generation, equipment, construction, first power, water and financing. Every row carries a link to its source, a source tier (primary, secondary, tertiary) and a confidence level. The build applies published lead-time assumptions to those rows to produce the traffic lights, the implied first-power date and the feasibility items. The rules are on the Method page; every change to the data appears on the Updates page, dated.

Where a plan has been announced but no completion reported, we may enter a placeholder date so a campus can be compared with others. Placeholders are marked low confidence and say so in their note. We never silently edit a published row; corrections are new rows.

Data & sources

  • Epoch AI, "AI data centers" (epoch.ai/data/ai-data-centers), CC BY 4.0. Campus list, satellite-based build-out notes and IT-power estimates for the campuses it covers. Attribution is shown on each campus page. Epoch AI has not reviewed or endorsed this site.
  • GeoNames (cities500), CC BY 4.0, for city coordinates; Natural Earth, public domain, for the map outline.
  • Public reporting, filings and announcements. We store facts, dates and a link; we do not reproduce article text beyond short attributed quotations. Each source remains the property of its publisher. If you are a rights holder and believe something here exceeds fair use, write to us and we will act on it.
  • Company names and marks (operators, utilities, equipment makers) belong to their owners and are used only to identify the projects and products discussed.
  • Our own additions: the hand-entered campus records, step tables, lead-time assumptions, equipment and utility tables, scores and charts are AI Loadmap's work, published under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). You may copy, redistribute and adapt them for any purpose, commercial included, by hand or by automated collection, as long as you credit "AI Loadmap (ailoadmap.com)", link to the page or chart you used, and say if you changed anything. Fields from Epoch AI need Epoch AI's credit as well. The licence covers our data and charts, not the source articles we link to, not company names or marks, and not the AI Loadmap name and logo as a brand.
  • How to cite: AI Loadmap, "{page or chart title}", data version {week}, {URL}. Every chart and campus page has a Cite button that fills this in.

Contact

To report an error, request a correction, raise a rights concern or ask about the data, write to contact@ailoadmap.com. For corrections, include the campus name, the row you mean and the source that supports the change. We read everything and answer what we can.

AI Loadmap is independent. It has no commercial relationship with any company, utility or agency named on the site.

See also the terms of use.