Supercomputers/Systems/Data Section Osaka B200 Supercluster
Japan · planned · mixed
Data Section Osaka B200 Supercluster
Operated by Data Section .
Researched with AI assistance, checked by scripts. These rows were researched by AI agents reading public operator, laboratory and vendor sources, appended by scripts and checked by automated validation; no subject-matter expert has reviewed them (how this site is made). A mechanical second-reader pass has since fetched every cited source: 1350 of 1540 systems have a readable citation that names them, and 64 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
- -
- Rpeak
- -
- Rmax ÷ Rpeak
- -
- Cores
- -
- Power
- -
- Per watt
- - 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 | - GIGA Computing GPU-enabled servers, 625 servers | Giga Computing | 625 reported | Reported | datasection.co.jp |
| Accelerator | NVIDIA B200 SXM Accelerators CUDA (Blackwell) · TSMC 4NP · 1,000 W Servers equipped with NVIDIA B200 GPUs (5,000 units / 625 servers), NVIDIA HGX B200 | NVIDIA | 5,000 reported | Reported | datasection.co.jp |
AI datacenter metrics
Figures beyond an accelerator count, each with who stated it. What can and cannot be compared across AI datacenters.
| Metric | Value | As of | Basis | Note | Source |
|---|---|---|---|---|---|
| Accelerators | 5,000 accelerators | 2025-07 | Operator stated | Datasection release: servers with 5,000 B200 GPUs (625 servers) procured for the Osaka supercluster | datasection.co.jp |
The read
A planned GPU-as-a-service supercluster in Osaka. Datasection's 4 July 2025 release says it had procured servers carrying 5,000 NVIDIA B200 GPUs (625 servers, which is eight per server) and would build the cluster to give over 100 ExaFLOPS of FP4 compute, a vendor-stated AI figure not recorded as FLOPS. It says the build responds to a request from a potential client described only as one of the world's largest cloud service providers, names GIGA Computing as server supplier and mentions its own TAIZA software for utilisation of large GPU clusters. The release gives no building, power figure, delivery or go-live date, so the status stays planned as of the release date; I found no later page confirming operation. The only source is the operator's own PDF. Compare sakura-koukaryoku-h200-container for another Japanese GPU-cloud buildout of about 1,000 GPUs.
Timeline
- 2025-07 Announced source
Change history
- 2026-10-17 Added Data Section Osaka B200 supercluster (planned, 5,000 GPUs)
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 |
|---|---|---|
| datasection.co.jp | names system only | - |
Checked 2026-10-18 by pnpm verify. Re-run it and the table changes with the web.