AI Data Center Buildout in 2023

This timeline slice groups facilities tied to 2023 through the best current public milestone year in the dataset. It captures 4.7 GW of disclosed capacity across this cohort.

Facilities
26
Tracked Capacity
4.7GW
Countries
17
Operators
30
Facility Capacity
Microsoft — Australia AI Expansion
Sydney, Melbourne, Canberra, Australia
1000MW
Oracle AI Cloud Infrastructure (US)
Hillsboro, United States
800MW
Microsoft Azure — Poland Central (Warsaw)
Warsaw, Poland
500MW
Tesla Dojo Supercomputer — Buffalo New York
Buffalo, United States
500MW
Meta — Eagle Mountain, Utah AI Data Center Campus
Eagle Mountain, United States
500MW
Tencent Tianjin AI Data Center (China)
Tianjin, China
300MW
Microsoft Azure — Italy North Region (Milan)
Milan, Italy
150MW
DeepSeek — AI Training Cluster
Hangzhou, China
100MW
NEBIUS AI Cloud — Mäntsälä Finland Data Center
Mäntsälä, Finland
100MW
IBM — Watson AI Data Centers (US)
Raleigh, United States
100MW
Microsoft Azure — Portugal Region (Lisbon)
Lisbon, Portugal
80MW
Amazon Web Services — Canada West Region (Calgary)
Calgary, Canada
80MW
Amazon Web Services — Melbourne Australia Region (ap-southeast-4)
Melbourne, Australia
80MW
Amazon Web Services — Israel Region (il-central-1, Tel Aviv)
Tel Aviv, Israel
60MW
Microsoft — Inflection AI Pi Infrastructure (Midwest)
Chicago, United States
50MW
Mistral AI — Training Cluster (OVHcloud, France)
Paris, France
40MW
Scaleway — H100 AI Cloud Data Center (France)
Vitry-sur-Seine, France
40MW
Google Cloud — New Zealand Region (Auckland)
Auckland, New Zealand
40MW
Google Cloud — Colombia AI Region
Bogotá, Colombia
40MW
Google Cloud — Argentina Region (Buenos Aires)
Buenos Aires, Argentina
40MW
Cerebras Systems — Condor Galaxy AI Supercomputer
Dallas, United States
30MW
SAP AI Core — Walldorf Data Center
Walldorf, Germany
30MW
UK AI Research Resource (AIRR) — Isambard-AI & Dawn Supercomputers
Bristol / Cambridge, United Kingdom
20MW
MareNostrum 5 Supercomputing Center
Barcelona, Spain
10MW
NVIDIA “Israel-1” AI Supercomputer (Israel)
Israel (cloud access), Israel
AIRAWAT AI Supercomputer (India)
Pune, India

Timeline Method

This page currently groups projects by the best single public milestone year in the dataset, stored as start_year. As the index fills out separate announcement, groundbreaking, and go-live fields, the year taxonomy can split into more precise public timeline surfaces and filtered feeds.