Sovereign AI
Sovereign programme record · Switzerland · 2024
Switzerland: ~$118M announced · disbursement not on record
Switzerland: Alps Supercomputer + Swiss AI Initiative
- Announced~$118M
- Government funding~$118M
- Committed—
- Disbursed—
Operating compute / delivered hardware. Reported operational or delivered: 10,752 NVIDIA Grace Hopper GH200 (physical hardware).
Hardware on record
| Model | Count | Status | Organisation / role |
|---|---|---|---|
| NVIDIA Grace Hopper GH200 | 10,752 | Operational | CSCS, an ETH Zurich institution |
Physical hardware: Alps’ Grace Hopper superchips at its September 2024 inauguration. Primary source · 2024-09-14 | |||
Empanelled partners · 6
| Partner | Accelerators | Model | Stage | Source |
|---|---|---|---|---|
| Swiss National Supercomputing Centre (CSCS) | 10,752 | NVIDIA GH200 Grace Hopper Superchip | Commissioned | Primary source · 2024-09-14 |
Physical hardware: Alps’ Grace Hopper superchips at its September 2024 inauguration. Primary source · 2024-09-14 Operates Alps in Lugano. ETH Zurich reported 10,752 NVIDIA Grace Hopper superchips at the inauguration on 14 September 2024. | ||||
| ETH Zurich | — | — | Empanelled | Primary source · 2023-12-01 |
Co-launched the Swiss AI Initiative with EPFL in December 2023; its CSCS supercomputing centre provides access to Alps.
| ||||
| EPFL | — | — | Empanelled | Primary source · 2023-12-01 |
Co-leads the Swiss AI Initiative with ETH Zurich through their respective AI centres.
| ||||
| Hewlett Packard Enterprise | — | — | Awarded | Primary source · 2024-09-14 |
Supplied the HPE Cray EX system used for Alps.
| ||||
| Paul Scherrer Institute | — | — | Empanelled | Primary source · 2024-09-14 |
Uses Alps through its virtual-cluster architecture for scientific research.
| ||||
| MeteoSwiss | — | — | Empanelled | Primary source · 2024-09-14 |
Uses Alps for higher-resolution weather forecasting in Switzerland.
| ||||
External dependence
NVIDIA hardware: High · US technology: High. These are editorial assessments based on the programme’s procurement disclosures.
Named partners: CSCS, ETH Zurich, EPFL, HPE, NVIDIA.
CSCS operates Alps in Lugano, with 10,752 NVIDIA Grace Hopper superchips reported at its September 2024 inauguration. The system supports the Swiss AI Initiative, led by ETH Zurich and EPFL. Their multilingual Apertus models were released on 2 September 2025 for public download. Alps also serves scientific workloads, including weather forecasting for MeteoSwiss and research at the Paul Scherrer Institute.
Data-quality caveats on this record · 3
- The GPU count is the 10,752 Grace Hopper superchips reported by ETH Zurich at the Alps inauguration on 14 September 2024.
- SWI swissinfo reported Alps’ capital cost as CHF100M (approximately US$118M) on 3 October 2024, funded from the public purse.
- The ETH Board allocated CHF 20M to Swiss AI Initiative research for 2025–2028.
The evidence trail
11 cited sources- Alps inauguration announcement (CSCS) — Primary source · accessed 2026-04-07
- ETH Zurich: Alps inauguration and hardware specification — Primary source · accessed 2026-03-28
- CSCS Alps technical specifications — Primary source · accessed 2026-03-28
- CHF 100M capital, CHF 37M operating budget — Press or analyst · accessed 2026-03-28
- Swiss AI Initiative launch — Primary source · accessed 2026-03-28
- Apertus release plans, July 2025 — Primary source · accessed 2026-03-28
- DC capacity target: main facility in Lugano; CSCS extending Alps infrastructure across Switzerland — Research database · no accessed date on record
- CSCS (Swiss National Supercomputing Centre): operator — Primary source · no accessed date on record
- CSCS: Apertus model release, 2 September 2025 — Primary source · accessed 2026-09-08
- SWI swissinfo: reported Alps capital cost and CSCS director interview — Press or analyst · accessed 2026-09-08
- ETH Zurich, Globe 4/2024, page 32: CHF 20M Swiss AI Initiative funding for 2025–2028 — Primary source · accessed 2026-09-08
Machine accessThe same record, machine-readable: the evidence-trail feed and the derivation-chain export.