Machinepower Index / Dossiers / Estonia
Rank 29–46 across the scoring rules tested. How this is tested
Estonia
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.
Epoch AI records no large AI clusters or data centres here yet.
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 | 5.6 |
| Firm Power Capacity | Dataset | 2.0 | |
| Sovereign Compute | Assessed | 2.0 | |
| Hardware & Silicon Access | Dataset | 4.6 | |
| Weights | Frontier Model Output | Dataset | 3.0 |
| Alignment Capability | Assessed | 5.0 | |
| Regulatory Standing | Dataset | 9.0 | |
| Talent Density | Assessed | 3.1 | |
| Will | Policy Capacity | Dataset | 9.6 |
| Public Sector Adoption | Dataset | 10.0 | |
| Workforce Transition | Dataset | 9.4 | |
| Public Trust | Assessed | 6.4 |
Assessed marks a measurement we judged rather than took from a dataset: 5 of the 12 here.
Sources
Permits are quick, but the whole country generates about 6 TWh a year.
View source →Named its AI Act authorities after the deadline; no safety statements or safety body.
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 →Feeds firm power, held at the 2.0 floor.
View source →No system on the June 2026 TOP500 and no compute controlled by the state, universities or firms in Epoch AI; set at the floor.
View source →Stanford does not count Estonia. Estimated from 0.1 million GitHub developers and Oxford Insights' diffusion pillar, averaged with an estimate from LinkedIn's AI talent share (1.23% of members), calibrated on the nations Stanford counts.
View source →Feeds governance and public-sector adoption; policy capacity is listed separately.
View source →Ipsos does not cover it. Mapped to the Ipsos scale through the 28 countries both surveys cover.
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 Estonia below its Potential. Its two lowest scores cost it the most: firm power capacity at 2.0 (part of Watts) and sovereign compute at 2.0 (part of Watts).
Assessment
Estonia scores higher than any other country on putting AI into public services, on a grid that generates about 6 TWh a year — far too little for compute at scale.
The art of the possible
It does not need to host compute to benefit: it can rent it from neighbours and keep doing what it does best, deploying first. AI Leap, which puts AI tools in every secondary school, is the next test.
Strategic play — Diffusion Advantage
Stay first to deploy: buy compute abroad, keep public services the testing ground, and export the systems that work.
Leverage — the nearest position. No chokepoint. Its position is example: other governments copy Estonian systems, from X-Road to e-residency.
Open Estonia in the live Index → — every source behind every number, and the comparison tool.