Under $10B AI Infrastructure Investment
Smaller but still meaningful AI data center and compute platform commitments, often tied to a single market or growth stage.
Methodology Note
Comparable totals only sum rows already expressed in USD billions in the export. Rows disclosed solely in euros, yen, or other local currencies remain in the dataset and taxonomy, but are excluded from the USD subtotal until the site adopts a pinned conversion method.
| Commitment | Announced |
|---|---|
|
Amazon Web Services — AP-SOUTHEAST-1 Singapore Region
Singapore region infrastructure commitment across the AP-SOUTHEAST-1 footprint.
|
$9B |
|
SK Group–AWS Ulsan AI Data Center (South Korea)
Joint SK Group and AWS 1 GW-class Ulsan AI campus.
|
$5.1B |
|
AWS — Thailand Cloud Region (Asia Pacific)
Thailand region investment across three availability zones.
|
$5B |
|
AWS — Santiago AI Region
Santiago region with three AZs and waterless cooling design.
|
$4B |
|
Microsoft — Sweden Central AI Infrastructure
Sweden Central expansion with 20,000 GPUs across Gävle, Sandviken, and Staffanstorp.
|
$3.2B |
|
Microsoft — São Paulo AI Infrastructure (Expansion)
Brazil cloud and AI infrastructure expansion centered on São Paulo.
|
$2.7B |
|
Google — Pryor Creek, Oklahoma AI Data Center
AI-driven Pryor campus expansion for Gemini, Vertex AI, and Google Cloud.
|
$2B |
|
Nvidia & Foxconn — Kaohsiung AI Factory
Taiwan sovereign AI factory led by Foxconn and NVIDIA.
|
$1.4B |
|
Microsoft — Greece AI Region (Athens)
Three-site Athens region centered on Spata and Koropi.
|
$1.1B |
|
Google — Chonburi (Thailand) AI Data Center
Thailand data center and cloud region investment anchored in Chonburi.
|
$1B |
|
Microsoft Azure — Poland Central (Warsaw)
Poland Central region rollout and national AI access expansion.
|
$1B |
|
Nebius & Mega Or — Israel AI Hub
Combined 80 MW Beit Shemesh and Masmiyya AI hub.
|
$880M |
|
CloudHQ — Querétaro AI Mega Campus
Initial six-building Querétaro campus commitment.
|
$600M |
|
Frontier Supercomputer — Oak Ridge National Laboratory
DOE exascale system delivered by HPE and AMD.
|
$600M |
|
Equinix — SG6 Singapore AI Data Center
Nine-story AI-ready SG6 expansion in Jurong.
|
$260M |
Browse Categories
Cloud regions and AI campuses funded directly by hyperscalers expanding training and inference capacity.
National and public-interest compute programs where governments or sovereign industrial policy anchor the buildout.
Colocation, GPU cloud, and developer-led projects built to capture external AI demand.
Public research systems and exascale supercomputers that also function as national AI training infrastructure.