Machinepower Index / Dossiers / Vietnam
Rank 25–32 across the scoring rules tested. How this is tested
Vietnam
Diffusion Advantage — For nations that cannot build the frontier but can use it faster than anyone: import the models and win on deployment, across the state and the economy, with public consent.
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: Vietnam controls 1,000 H100-equivalents, wherever the sites are. 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 | 6.2 |
| Firm Power Capacity | Dataset | 5.5 | |
| Sovereign Compute | Dataset | 5.0 | |
| Hardware & Silicon Access | Dataset | 4.8 | |
| Weights | Frontier Model Output | Dataset | 3.0 |
| Alignment Capability | Assessed | 3.6 | |
| Regulatory Standing | Dataset | 8.8 | |
| Talent Density | Assessed | 5.3 | |
| Will | Policy Capacity | Dataset | 5.4 |
| Public Sector Adoption | Dataset | 7.3 | |
| Workforce Transition | Dataset | 5.4 | |
| Public Trust | Assessed | 7.7 |
Assessed marks a measurement we judged rather than took from a dataset: 4 of the 12 here.
Sources
Data centres may now be wholly foreign-owned and buy power directly; the northern grid ran short in 2023.
View source →Signed none of the eight international AI commitments tracked; AI rules sit inside a 2025 digital-industry law.
View source →A cross-check for planning and permitting, which stays assessed: it scores permits and electricity connections for ordinary firms, not grid-scale connections for data centres.
View source →Largest: FPT AI Factory.
View source →Stanford does not count Vietnam. Estimated from 3.0 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 →No AI survey in the last three years covers it; set at the median of the 15 surveyed emerging and developing economies in the Index.
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 →Half of public trust: support for building more data centres, scored as 1 + share ÷ 10 and averaged with trust in AI.
View source →Bottlenecks
These hold Vietnam below its Potential. Its two lowest scores cost it the most: frontier model output at 3.0 (part of Weights) and alignment capability at 3.6 (part of Weights).
Assessment
Vietnam builds fast and has two national champions, FPT and Viettel, running AI factories on Nvidia hardware. What it lacks is research depth and a model of its own.
The art of the possible
Its assembly base and three million developers give it a way up the chain if FPT and Viettel turn their GPU clusters into services the region buys.
Strategic play — Diffusion Advantage
Grow the champions into regional suppliers: let FPT and Viettel sell compute and AI services across Southeast Asia, and use the assembly base to bargain for chips.
Leverage — the nearest position. No chokepoint. Vietnam assembles and tests chips for others, and is one of five states to sign both the US-led AI statement and China's WAICO; neither gives it a lever.
Open Vietnam in the live Index → — every source behind every number, and the comparison tool.