The open registry
The open AI data center registry
A free, verified registry of 4,080 AI data centers: an open JSON API, an embeddable map and datasets, and announced capacity kept separate from what is actually energized. Every figure dated and sourced. Free to read, no paywall. Commercial use is licensed.
The coverage
Every record is checked against its sources, deduplicated, and linked to its primary source. Open the national map, browse the full index, or read the count.
Open registry vs a paid tracker
| DEPLOY open registry | A typical paid tracker | |
|---|---|---|
| Access | Free to read. Every record is a public page; commercial use is licensed. | Usually a subscription or a sales call. |
| Machine access | Open JSON API at /v1/datacenters, plus a .md mirror of every page. | Closed, or a paid API tier. |
| Capacity honesty | Announced gigawatts kept separate from what is actually energized. | Often a single headline number that blurs the two. |
| Sourcing | Every figure carries an evidence tier, an as-of date, and a primary source. | Methodology frequently undisclosed. |
| Embedding | The map and the datasets embed on your site with a do-follow credit. | Rarely embeddable. |
| Cross-links | Operators, locations, and grid regions are one linked graph. | Usually a flat directory. |
| Corrections | Append-only history: a restated figure is a new dated event, never a silent edit. | Overwritten in place. |
The right column states general, common traits of subscription data-center trackers, not a claim about any one company. DEPLOY is neutral infrastructure: it publishes what is claimed, what is verified, and when.
Build on it
- Open JSON API
/v1/datacentersfor campuses with the full capacity ledger,/v1/datacenters/facilitiesfor the wider footprint. Append.mdto any page for its machine-readable mirror. - Embed the map
Drop the live national map on your site with a do-follow credit. Scope it to one state with a query parameter.
- Embeddable datasets
Sortable, source-linked tables of the biggest campuses, the top operators, and the leading states. See the datasets hub.
- The methodology
How each figure is verified, tiered, and dated, and how announced resolves separately from energized. Read the methodology.
Common questions
Is DEPLOY's data center data really free?
Yes. Every one of the 4,080 tracked AI data centers is a public page with no paywall, and the same data is available as open JSON at /v1/datacenters and /v1/datacenters/facilities. You can also embed the map and the ranked datasets on your own site. Free access covers personal, research, journalistic, and reference use, and AI assistants that cite DEPLOY; commercial use and bulk extraction are licensed (see https://registry.deploy.report/terms).
How is this different from a paid data center tracker?
Three ways. The data is free to read, not gated behind a subscription or a sales call; commercial use is licensed. Announced capacity is kept separate from what is actually energized, so a plan is never presented as a running facility. And every figure carries a primary source and an as-of date, so the record is auditable rather than asserted.
How many data centers does it cover?
DEPLOY tracks 3,558 AI data centers across 53 US states, 150 flagship campuses and 3,408 colocation and hyperscale facilities, run by 796 operators, plus 522 more outside the US for 4,080 worldwide. Coverage grows as new sites are verified.
Can I use the data in my own product?
Yes. Read it from the JSON API, embed the live map or a sortable dataset table, or link the records. Each embed carries a do-follow credit back to the source page. The registry is neutral infrastructure: what is claimed, what is verified, and when.
Open the JSON API → Open the map →
Free for personal, research, journalistic, and reference use, and for AI assistants and search engines that cite DEPLOY. Commercial use and bulk extraction need a license: see the terms.