Asia
Thailand
5 AI supercomputers outside the tracked campuses are in Epoch AI’s list to Jul 2025, with 2,270 H100-equivalents. 5 public cloud regions are live here, from 5 providers.
- AI campuses
- None tracked
AI supercomputers
Epoch AI · to Jul 2025| Name | Owner | H100e | Status | Since |
|---|---|---|---|---|
| GreenNode Bangkok Cluster | VNG Corporation | 1,024 | Existing | Jun 2024 |
| SIAM AI HGX | SIAM AI | 1,024 | Existing (likely) | Sep 2024 |
| NSTDA Supercomputer Center (ThaiSC) LANTA | National Science and Technology Development Agency (NSTDA) | 222 | Existing | Nov 2022 |
| Alibaba Cloud in Thailand Phase 1 | Alibaba | — | Existing | Jul 2022 |
| Alibaba Cloud in Thailand Phase 2 | Alibaba | — | Existing | Jan 2025 |
Cloud regions
6 providers · 6 regionsAWS
- Asia Pacific (Thailand) ap-southeast-7 Bangkok Live
Microsoft Azure
- new datacenter region in Thailand Bangkok Announced
Google Cloud
- Bangkok (asia-southeast3) asia-southeast3 Bangkok Live
Alibaba Cloud
- Thailand (Bangkok) ap-southeast-7 Bangkok Live
Tencent Cloud
- 亚太和中东(曼谷) ap-bangkok Bangkok Live
Huawei Cloud
- AP-Bangkok ap-southeast-2 Bangkok Live
Gas made 68% of the power in Thailand
share of generation · Ember · Jul 2025 to Jun 2026Thailand World
- Gas 68% against 21% for World
- Coal 14% against 33% for World
- Other fossil <1% against 3% for World
- Nuclear 0% against 9% for World
- Wind 2% against 9% for World
- Solar 6% against 10% for World
- Hydro 5% against 14% for World
- Bio & other 5% against 2% for World
Fossil 83% (−1 pt on a year earlier) · nuclear 0% · renewables 17% · 533 g CO2e per kWh, life cycle · Jul 2025 to Jun 2026. Generation only, without imports. Every country in the registry, side by side.
Data-centre electricity
International Energy AgencyAsia Pacific excluding China, as the IEA defines it
- 2030, base case around 100 TWh IEA 2026, p. 64
Regionally, the Asia Pacific region – excluding China – is projected to consume around 100 TWh for data centres in 2030, which accounts for 11% of future global electricity consumption by data centres.
Sources
Every dataset → Status → Glossary →- Ember, Monthly Electricity Data and Yearly Electricity Data, CC BY 4.0, retrieved 11 Oct 2026. Shares and intensity computed here.
- Epoch AI, Data on GPU clusters, CC BY 4.0, retrieved 16 Oct 2026. Deprecated by Epoch AI; no longer updated.
- Cloud providers, Public region lists of AWS, Microsoft Azure, Google Cloud, Oracle Cloud Infrastructure, Alibaba Cloud, Tencent Cloud, Huawei Cloud, IBM Cloud, CoreWeave, Nebius, Facts from each provider’s public region list, retrieved 11 Oct 2026.
- OpenStreetMap contributors, Nominatim geocoding of cloud region cities, ODbL 1.0, retrieved 11 Oct 2026.
- International Energy Agency, Energy and AI (2025); Key Questions on Energy and AI (2026); Energy and AI in East Asia (2026), CC BY 4.0, retrieved 11 Oct 2026.