Machinepower Index / Dossiers / Thailand
Rank 26–29 across the scoring rules tested. How this is tested
Thailand
Compute Rentier — For nations whose main asset is the ground itself, meaning cheap clean power, land and stability: host other nations' compute and turn geography into rent, jobs and relevance.
How the Score, Potential and bottleneck cost are worked out: methodology.
The three layers
Energy, grids, data centres and silicon — the physical capacity to run models at scale.
Compute tracked by Epoch AI: Thailand controls 1,200 H100-equivalents, wherever the sites are; hosts 1,000 for foreign firms. Hosted compute is shown, not scored.
Models, alignment capability, standards influence and the talent that produces them.
The capacity to wire AI into the state and the economy while holding public consent.
Where the score comes from
| Layer | Sub-dimension | Basis | Score |
|---|---|---|---|
| Watts | Planning & Permitting | Assessed | 6.6 |
| Firm Power Capacity | Dataset | 5.1 | |
| Sovereign Compute | Dataset | 3.5 | |
| Hardware & Silicon Access | Dataset | 4.0 | |
| Weights | Frontier Model Output | Dataset | 3.0 |
| Alignment Capability | Assessed | 3.8 | |
| Regulatory Standing | Dataset | 7.0 | |
| Talent Density | Assessed | 5.0 | |
| Will | Policy Capacity | Dataset | 8.8 |
| Public Sector Adoption | Dataset | 8.6 | |
| Workforce Transition | Dataset | 6.4 | |
| Public Trust | Dataset | 8.4 |
Assessed marks a measurement we judged rather than took from a dataset: 3 of the 12 here.
Sources
Approves data centres at speed but paused new promotion applications in April 2026 to write power and water rules.
View source →A draft AI law and official guidance; no safety body or safety statements.
View source →Largest: PTTEP's THE CRUST.
View source →Stanford does not count Thailand. Estimated from 1.6 million GitHub developers and Oxford Insights' diffusion pillar, calibrated on the nations Stanford counts.
View source →Feeds governance and public-sector adoption; policy capacity is listed separately.
View source →The Index scores public trust as 1 + share ÷ 10. Survey of 23,532 adults across 32 countries, March to April 2026.
View source →Feeds workforce, rescaled across 173 economies.
View source →Fossil, nuclear, hydro and other dispatchable plant; averaged with generation for firm power.
View source →Bottlenecks
These hold Thailand below its Potential. Its two lowest scores cost it the most: frontier model output at 3.0 (part of Weights) and sovereign compute at 3.5 (part of Watts).
Assessment
Thailand has approved more data-centre investment in a year than most countries see in a decade, led by TikTok. It controls little of that compute, and in April 2026 paused new approvals to write rules on power and water.
The art of the possible
Siam AI, a Thai Nvidia cloud partner, shows some capacity can stay national. Tying new campuses to grid charges and local access would make hosting fund home-grown capability.
Strategic play — Compute Rentier
Rent the ground on terms: price power and water properly under the new rules, and reserve a share of hosted capacity for Thai firms and the state.
Leverage — the nearest position. No chokepoint. Its position is a well-connected host between the American and Chinese stacks: TikTok and the US hyperscalers are both building there.
Open Thailand in the live Index → — every source behind every number, and the comparison tool.