Machinepower Index / Dossiers / Switzerland
Rank 13–22 across the scoring rules tested. How this is tested
Switzerland
Specialist Power — For nations among the world's best at something the frontier needs but could find elsewhere: keep earning the lead before others catch up, and trade expertise for access.
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: Switzerland controls 17,000 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 | 5.0 |
| Firm Power Capacity | Dataset | 3.8 | |
| Sovereign Compute | Dataset | 6.7 | |
| Hardware & Silicon Access | Dataset | 4.7 | |
| Weights | Frontier Model Output | Assessed | 4.0 |
| Alignment Capability | Assessed | 6.5 | |
| Regulatory Standing | Dataset | 8.9 | |
| Talent Density | Assessed Sources differ | 7.8 | |
| Will | Policy Capacity | Dataset | 3.9 |
| Public Sector Adoption | Dataset | 7.4 | |
| Workforce Transition | Dataset | 9.5 | |
| Public Trust | Dataset Sources differ | 6.1 |
Assessed marks a measurement we judged rather than took from a dataset: 4 of the 12 here.
Sources differ marks a cell where two sources disagree by more than a point. Talent Density: Judged at 7.8; both data sources put it lower (5.8 and 5.0). Public Trust: Ipsos and Melbourne/KPMG differ by more than a point (5.5 against 6.6); the cell is their average.
Sources
Apertus, a fully open multilingual model released in September 2025 by EPFL, ETH Zurich and the national supercomputing centre, is a credible public model at 8B and 70B parameters, well below the frontier.
View source →Alps at CSCS Lugano is the tenth-fastest system on Earth on the June 2026 list; Switzerland's ~450 PF aggregate is remarkable for a nine-million-person state.
View source →The validated January 2026 report rates Swiss governance capacity at 89.20/100, with public-sector adoption at 73.93.
View source →Used for the workforce measurement. The IMF combines schooling, pay, labour mobility and active labour-market policies; the Index rescales it so the weakest of 173 economies scores 0 and the strongest 10.
View source →Fossil, nuclear, hydro and other dispatchable plant; averaged with generation for firm power.
View source →Second source for public trust, averaged with Ipsos.
View source →Bottlenecks
These hold Switzerland below its Potential. Its two lowest scores cost it the most: firm power capacity at 3.8 (part of Watts) and policy capacity at 3.9 (part of Will).
Assessment
The Alps machine at CSCS is a top-ten system on Earth and Apertus is among Europe's most credible sovereign open models — a remarkable intelligence layer for nine million people. But Switzerland runs it all on 65 TWh and a public among the most AI-sceptical surveyed (45%). Precision, not scale, is the Swiss condition.
The art of the possible
Switzerland cannot out-build anyone; it can out-trust everyone. The play is to become the neutral ground of the AI age — sovereign evaluation, model auditing and data escrow for states that trust neither Washington nor Beijing. The Geneva of weights, priced accordingly.
Strategic play — Specialist Power
Out-trust everyone it cannot out-build: become the neutral ground for evaluation, model auditing and data escrow, the place rivals park the models they will not host in each other's clouds, and price that neutrality accordingly.
Leverage — the nearest position. Neutrality as a position: a top-ten public machine and a jurisdiction both machinepowers trust more than each other.
Open Switzerland in the live Index → — every source behind every number, and the comparison tool.