Machinepower Index / Dossiers / Saudi Arabia
Rank 9–14 across the scoring rules tested. How this is tested
Saudi Arabia
Sovereign Capital — For nations with more money than research depth: buy the frontier rather than build it, acquiring chips, talent, data centres and equity stakes at scale while the window is open.
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: Saudi Arabia controls 8,400 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 | 9.2 |
| Firm Power Capacity | Dataset | 6.0 | |
| Sovereign Compute | Dataset | 6.0 | |
| Hardware & Silicon Access | Dataset | 4.0 | |
| Weights | Frontier Model Output | Assessed | 3.6 |
| Alignment Capability | Assessed | 5.8 | |
| Regulatory Standing | Dataset | 9.3 | |
| Talent Density | Dataset | 4.6 | |
| Will | Policy Capacity | Dataset | 9.2 |
| Public Sector Adoption | Dataset | 8.9 | |
| Workforce Transition | Dataset | 8.9 | |
| Public Trust | Assessed | 8.2 |
Assessed marks a measurement we judged rather than took from a dataset: 4 of the 12 here.
Sources
HUMAIN's own ALLaM family tops out at 34B parameters and is an Arabic specialist, absent from the top tiers of the main independent leaderboards. Its largest model, the 428B humain-m3, was commissioned from China's MiniMax and built on its model line.
View source →NVIDIA's May 2025 announcement commits up to 500 megawatts of AI infrastructure with HUMAIN, first phase 18,000 GB300 systems. HUMAIN separately stated a 600,000-GPU ambition at the November 2025 US-Saudi Investment Forum.
View source →The PIF-owned national AI company consolidates data centres, cloud, and model development under a single state-backed vehicle with multi-campus construction underway.
View source →The Saudi Data & AI Authority coordinates the national strategy; top-down mandates deliver exceptional deployment speed, while the domestic research base remains the thinnest among Tier-2 peers.
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 talent, averaged with Stanford's count.
View source →Capital is shown beside the Score, not scored.
View source →Bottlenecks
These hold Saudi Arabia below its Potential. Its two lowest scores cost it the most: frontier model output at 3.6 (part of Weights) and hardware & silicon access at 4.0 (part of Watts).
Assessment
Saudi Arabia is converting energy and capital abundance into compute at extraordinary speed through HUMAIN, its PIF-backed sovereign AI champion, and multi-hundred-thousand-GPU agreements with US suppliers. Its constraint is the intelligence layer: a thin domestic research base means the hardware build-out currently runs ahead of the talent to exploit it.
The art of the possible
Saudi Arabia is executing the fastest energy-to-compute conversion in history; its upside hinges on the talent layer catching up. If KAUST-anchored research and international recruitment scale with the GPU build-out, HUMAIN becomes a frontier lab rather than a hosting company.
Strategic play — Sovereign Capital
Use the window while sellers still need buyers: convert energy and capital into GPU agreements quickly, then make KAUST-anchored research and international recruitment keep pace, so HUMAIN becomes a frontier lab rather than a hosting company once the datacentres are full.
Leverage — the nearest position. The same play as the UAE at larger scale: energy and capital converted into GPU agreements that make it a swing buyer in a supply-constrained market. The lever belongs to the seller until the datacentres are full.
Open Saudi Arabia in the live Index → — every source behind every number, and the comparison tool.