Supercomputers/Systems/GMO GPU Cloud
Japan · operational · ai training
GMO GPU Cloud
Operated by GMO Internet Group .
Phase 1 seed dataset, compiled by hand. These rows were built from public operator, laboratory and vendor sources. A mechanical second-reader pass has since fetched every cited source: 1092 of 1153 systems have a readable citation that names them, and 40 are genuinely weakly sourced. Every claim carries its source and a confidence tier. Treat anything below verified as a lead, not a citation.
Measured performance
- Rmax
- 10.05 PFlop/s
- Rpeak
- 17.36 PFlop/s
- Rmax ÷ Rpeak
- 57.9%
- Cores
- 35,328
- Power
- 186.78 kW
- Per watt
- 53.8 GF/W
Figures are as last publicly reported for the configuration described below, not a live measurement. Where a system was upgraded in place, the post-upgrade configuration is shown and the earlier one appears in the timeline.
What this machine is made of
One row per supplier relationship. “Supplier at build” is the company that shipped the part at the time; where that company has since been acquired, the parent it rolls up to today is shown beside it. That distinction is what makes ticker-level aggregation possible across a thirty-year dataset.
| Role | Part | Supplier at build | Quantity | Confidence | Source |
|---|---|---|---|---|---|
| Integrator | - Dell PowerEdge XE9640 | Dell Technologies | - | Reported | top500.org |
| Accelerator | NVIDIA H200 SXM Accelerators CUDA (Hopper GH100) · 16,896C · 1.98 GHz · TSMC 4N · 700 W 256 NVIDIA H200 GPUs across 32 server nodes | NVIDIA | 256 reported | Reported | internet.gmo |
| CPU | - Xeon Platinum 8480L 56C 2GHz | Intel | - | Reported | top500.org |
| Interconnect | - NVIDIA Spectrum-X Ethernet | NVIDIA | - | Verified | internet.gmotop500.org |
The read
GMO Internet Group's GPU cloud service, launched in November 2024. GMO's June 2025 release states a 32-node, 256-GPU NVIDIA H200 configuration on NVIDIA Spectrum-X networking that measured 10.05 PFlop/s and 53.81 GFlops/Watt, ranking 34th on the June 2025 Green500 and first in Japan; TOP500 lists the same 10.05 PFlop/s, Rpeak 17.36 PFlop/s, 35,328 cores and 186.78 kW on a Dell PowerEdge XE9640 with Xeon Platinum 8480L hosts and Ubuntu 22.04.4, ranking it 141st in June 2025, 164th in November 2025 and 185th in June 2026. TOP500's page text names the accelerator as a Tesla GP100, which conflicts with GMO's H200, so the H200 comes from the operator. A larger 96-node, 768-GPU H200 run of 38.06 PFlop/s was 37th on the November 2024 TOP500 per GMO; that run is not the one carried in this row.
Timeline
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Change history
- 2026-10-02 Added GMO GPU Cloud, a Japanese H200 cloud
Source check
We fetched this system's own citations and recorded whether each page actually mentions it. This is a corroboration signal, not a fact check, and it is published so you can see how well the sourcing holds up rather than take it on trust.
| Cited source | Result | Found on the page |
|---|---|---|
| internet.gmo | names system, part and count | NVIDIA H200 SXM · counts: 256 |
| top500.org | names system and part | NVIDIA H200 SXM |
Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.
Inside a node
Latency-optimised versus throughput-optimised processors, NUMA, and the end of the host-to-device copy.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.