Supercomputers/Systems/Naver B200 4K Cluster
South Korea · operational · ai training
Naver B200 4K Cluster
Operated by Naver .
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
- -
- 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 | - Naver combined its in-house cooling, power-distribution and network-optimization technologies to build the cluster | Naver | - | Reported | koreatimes.co.kr |
| Accelerator | NVIDIA B200 SXM Accelerators CUDA (Blackwell) · TSMC 4NP · 1,000 W 4,000 Nvidia B200 GPUs | NVIDIA | 4,000 reported | Reported | koreatimes.co.krkoreatechtoday.com |
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 | 4,000 accelerators | 2026-01 | Operator stated | Naver reportedly completed a 4,000 B200 cluster in a leased Seoul facility (Jan 2026 press report). | koreatechtoday.com |
The read
Naver completed and announced a cluster of 4,000 Nvidia B200 GPUs on 8 January 2026 (Korea Times; a second outlet dates the report one day later), described as the largest of its kind in South Korea and built with Naver's in-house cooling, power-distribution and network optimization. Naver said a 72-billion-parameter model that took about 18 months on its earlier 2,048-GPU A100 infrastructure could be trained in roughly six weeks on the new cluster, and that its aggregate performance is comparable to high-ranking TOP500 systems; both are company claims and no HPL result is published. The two outlets disagree on siting (one says a leased Seoul facility, the other gives no location), so the site is left unlocated. Intended for Naver's next-generation Omni (text, image, video, audio) models.
Change history
- 2026-10-04 Added AI datacenter metrics for naver-b200-4k
- 2026-10-01 Added Naver B200 4K Cluster (Naver)
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 |
|---|---|---|
| koreatechtoday.com | names system, part and count | NVIDIA B200 SXM · counts: 4000 |
| koreatimes.co.kr | names system, part and count | NVIDIA B200 SXM · counts: 4000 |
Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.