Machinepower Index / Dossiers / New Zealand
Rank 36–43 across the scoring rules tested. How this is tested
New Zealand
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.
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 | 6.0 |
| Firm Power Capacity | Dataset | 3.0 | |
| Sovereign Compute | Assessed | 2.0 | |
| Hardware & Silicon Access | Dataset | 4.5 | |
| Weights | Frontier Model Output | Dataset | 3.0 |
| Alignment Capability | Assessed | 5.4 | |
| Regulatory Standing | Dataset | 7.2 | |
| Talent Density | Assessed | 4.0 | |
| Will | Policy Capacity | Dataset | 9.2 |
| Public Sector Adoption | Dataset | 6.4 | |
| Workforce Transition | Dataset | 9.1 | |
| Public Trust | Dataset | 5.4 |
Assessed marks a measurement we judged rather than took from a dataset: 4 of the 12 here.
Sources
Large projects can skip ordinary consents, and a 280 MW AI campus in Southland won approval in March 2026.
View source →A 2025 strategy with safety commitments and the Seoul statement; relies on partners for testing.
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 →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 New Zealand. Estimated from 0.4 million GitHub developers and Oxford Insights' diffusion pillar, averaged with an estimate from LinkedIn's AI talent share (0.70% of members), 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 →Second source for public trust, averaged with Ipsos.
View source →Bottlenecks
These hold New Zealand below its Potential. Its two lowest scores cost it the most: sovereign compute at 2.0 (part of Watts) and firm power capacity at 3.0 (part of Watts).
Assessment
New Zealand has clean hydro power, stable politics and fast-track consents, and almost no compute or models of its own. Its first large AI campus, in Southland, won consent in 2026.
The art of the possible
Southland's power is the prize: the Tiwai Point smelter uses about a tenth of the country's electricity, and the new campus would be the next-largest user. Five Eyes membership makes it a host allies trust.
Strategic play — Compute Rentier
Host allied compute on renewable power, and spend the rent on the research capacity New Zealand lacks.
Leverage — the nearest position. No chokepoint. Its position is trust and distance: a Five Eyes member with renewable power, far from any conflict.
Open New Zealand in the live Index → — every source behind every number, and the comparison tool.