How many AI data centers are tracked for Training?

Name: Training AI data center facilities
Creator: AI Data Center Index
License: https://aidatacenterindex.com/data/

The AI Data Center Index tracks 39 facilities for Training with 23.4 GW of selected tracked capacity.

AI Workload Focus

Training

AI data centers focused on Training workloads — where operators are building and at what scale. 39 facilities with 23.4 GW of selected tracked capacity.

Facility	Operator	Location	Capacity ↓	Status	Energy	Workload	Sources
Meta Hyperion AI Data Center (Louisiana)	Meta Entergy Louisiana	United States Richland Parish, Louisiana	5,000 MW	under construction; future 5 GW expansion not yet delivered	Nuclear	Training	4
Microsoft Project Fairwater (Mount Pleasant, Wisconsin)	Microsoft NVIDIA	United States Mount Pleasant, Wisconsin	3,300 MW	startup activity and equipment bring-up reported in May 2026; full campus buildout not complete	Other	Training	4
Stock Farm Road 3 GW AI Data Center (South Korea)	Stock Farm Road, Inc.Fir Hills Jeollanam-do Provincial Government	South Korea Jeollanam-do	3,000 MW	announced MOU-stage proposal; early-stage and not fully funded	Other	Training	3
xAI Colossus / MACROHARDRR Greater Memphis AI Campus	xAI	United States Greater Memphis / Southaven xAI region	2,000 MW	Colossus operational; Southaven MACROHARDRR expansion and power context still developing	Gas	Training	5
Vantage Frontier Shackelford County AI Campus	Vantage Data Centers	United States Shackelford County, Texas	1,400 MW	construction began before August 2025 announcement; first building scheduled for H2 2026	Other	Training	4
Meta-Oklo Southern Ohio Nuclear Power Campus	Meta Oklo Inc.	United States Pike County, Ohio	1,200 MW	agreement announced January 2026; pre-construction and site characterization slated for 2026	Nuclear	Training	2
Google / AdaniConneX Visakhapatnam AI Hub	Google Google Cloud AdaniConneX Adani Enterprises Nxtra by Airtel	India Andhra Pradesh / Visakhapatnam	1,000 MW	under construction; Google broke ground on April 28, 2026	Other	Training	3
Meta Prometheus AI Supercluster (New Albany, Ohio)	Meta	United States New Albany, Ohio	1,000 MW	New Albany campus active and expanding; Prometheus described as a gigawatt-scale 2026 AI supercluster buildout	Nuclear	Training	5
ByteDance — Johor AI Data Center	ByteDance (TikTok)	Malaysia Johor	500 MW	under construction	Other	Training	1
Meta — Eagle Mountain, Utah AI Data Center Campus	Meta	United States Utah	500 MW	under construction	Renewable	Training	2
Tesla Dojo Supercomputer — Buffalo New York	Tesla	United States New York	500 MW	operational	Mixed	Training	3
YTL & NVIDIA — Johor AI Cloud	YTL Power & NVIDIA	Malaysia Johor	500 MW	operational	Renewable	Training	1
Meta — Altoona, Iowa AI Data Center Campus	Meta	United States Iowa	350 MW	operational	Renewable	Training	2
Meta — Fort Worth, Texas AI Training Campus	Meta	United States Texas	350 MW	operational	Renewable	Training	2
Bulk Data Centers — N01 AI Campus	Bulk Data Centers	Norway Agder	300 MW	under construction	Other	Training	1
Digital Realty UK AI Mega-Campus	Digital Realty	United Kingdom London/Slough	300 MW	planned	Renewable	Training	1
Meta — Luleå, Sweden AI Data Center Campus	Meta	Sweden Norrbotten	300 MW	operational	Renewable	Training	2
EcoDataCenter — Borlänge AI Mega Campus	EcoDataCenter	Sweden Dalarna	250 MW	under construction	Renewable	Training	1
Stargate Norway — Narvik AI Gigafactory	OpenAI, Nscale & Aker	Norway Nordland	230 MW	under construction	Other	Training	1
SINES 4.0 — SIN02 AI Gigafactory	Start Campus	Portugal Alentejo	180 MW	under construction	Renewable	Training	1
Vantage Data Centers Quebec Expansion (Montreal & Quebec City)	Vantage Data Centers	Canada Quebec	178 MW	operational	Other	Training	1
Reliance / Meta Jamnagar AI Data Center	Reliance Industries Jio Meta	India Gujarat / Jamnagar	168 MW	planned built-to-suit AI data center; RIL says 168 MW to be delivered within two years from June 2026	Renewable	Training	3
Magnora — Hämeenlinna AI Data Center	Magnora & NEEG	Finland Kanta-Häme	120 MW	planned	Other	Training	1
Microsoft — Markham AI Megaproject	Microsoft	Canada Ontario	112 MW	under construction	Other	Training	1
DeepSeek — AI Training Cluster	DeepSeek High-Flyer Quant	China Zhejiang	100 MW	operational	Mixed	Training	3
Ethiopia — National AI Factory	Ethiopian AI Institute & International Investors	Ethiopia Addis Ababa	100 MW	planned	Other	Training	1
Naver — Chuncheon Data Center (Gak)	Naver	South Korea Gangwon	100 MW	operational	Renewable	Training	2
Waymo / Google — Autonomous Driving AI Training Cluster	Google	United States California	80 MW	operational	Renewable	Training	2
Yotta D2 / NVIDIA Blackwell AI Supercluster (Greater Noida)	Yotta Data Services NVIDIA	India Uttar Pradesh / Greater Noida	60 MW	under construction / planned AI supercluster; expected go-live by August 2026	Other	Training	3
Mistral Compute / Eclairion Bruyeres-le-Chatel AI Cluster	Mistral AI Eclairion Fluidstack NVIDIA	France Bruyeres-le-Chatel, Essonne	40 MW	operational ramp; Mistral Compute serving production while physical capacity scales	Other	Training	5
Naver — B200 4K AI Cluster	Naver Cloud	South Korea Sejong	40 MW	operational	Other	Training	1
Aurora Supercomputer — Argonne National Laboratory	Argonne National Laboratory Intel Hewlett Packard Enterprise	United States Illinois	39 MW	operational	Other	Training	6
El Capitan Supercomputer — Lawrence Livermore National Laboratory	Lawrence Livermore National Laboratory Hewlett Packard Enterprise AMD	United States California	30 MW	operational	Other	Training	6
Frontier Supercomputer — Oak Ridge National Laboratory	Oak Ridge National Laboratory Hewlett Packard Enterprise AMD	United States Tennessee	25 MW	operational	Other	Training	6
Scaleway / OpCore DC5 (PAR2) Data Center	Scaleway OpCore iliad Group	France Saint-Ouen-l'Aumone / Paris region	16 MW	operational data center	Renewable	Training	5
Microsoft West Des Moines AI Datacenter Cluster	Microsoft	United States West Des Moines, Iowa	—	operational West Des Moines datacenter cluster with ongoing expansions	Renewable	Training	3
NVIDIA Eos AI Supercomputer	NVIDIA	United States Location not disclosed	—	operational	Other	Training	4
Scaleway Nabu H100 AI Supercomputer (DC5/PAR2)	Scaleway iliad Group NVIDIA	France DC5 / PAR2, Paris region	—	operational GPU supercomputer / cloud cluster	Renewable	Training	4
SenseTime Lin-gang AIDC (China)	SenseTime	China Lin-gang Special Area, Shanghai	—	operational	Other	Training	2

Training

Capacity Pipeline

Data Quality

How this slice breaks down

Status distribution

Top countries by capacity

Energy mix

Quick Links

Machine-readable Data

Frequently asked questions

Operators Present

Countries Present

Status Mix

Facilities