Brazil · operational · ai training

Jairu

Also known as USP Jairu AI cluster; Jairu supercomputer

Operated by USP CIAAM at USP CIAAM (Sao Paulo) .

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	- Supermicro HGX B200 servers	Supermicro SMCINASDAQ	-	Reported	telesintese.com.br
Integrator	- Scherm Brasil named as integrator	Scherm Brasil	-	Reported	telesintese.com.br
Accelerator	NVIDIA B200 SXM Accelerators CUDA (Blackwell) · TSMC 4NP · 1,000 W 96 NVIDIA B200 SXM GPUs in 12 HGX 8x B200 nodes	NVIDIA NVDANASDAQ	96 reported	Reported	telesintese.com.br
CPU	- AMD EPYC processors on the five head nodes	AMD AMDNASDAQ	-	Reported	telesintese.com.br
Interconnect	NVIDIA InfiniBand NDR (Quantum-2) Interconnect Fat tree or dragonfly+ · 400 Gb/s per port · Under 0.6 microseconds InfiniBand NDR up to 800 Gb/s between nodes	NVIDIA NVDANASDAQ	-	Reported	telesintese.com.br

The read

The University of Sao Paulo's AI cluster for its Center for Artificial Intelligence and Machine Learning (CIAAM-USP), inaugurated on 13 February 2026 and described by USP as the largest in Latin America with Blackwell B200 GPUs. Press coverage lists 96 NVIDIA B200 SXM GPUs in 12 HGX 8x B200 compute nodes joined by NVLink, five AMD EPYC head nodes, InfiniBand NDR up to 800 Gb/s plus 200 Gb/s Ethernet, about 300 TB of BeeGFS storage and NVIDIA AI Enterprise with Base Command Manager, using Supermicro servers integrated by Scherm Brasil with Positivo Servers and Solutions as a partner. Procurement cost R$40 million (about US$8 million), imported directly from the United States and assembled within 30 days of the contract. No power or peak figure was found. The two outlets are independent but both relay USP's announcement, so the tier is reported.

Timeline

2026-02 First operational source

Site & facility

USP CIAAM (Sao Paulo)

Location: Sao Paulo, Brazil

telesintese.com.br inforchannel.com.br

Change history

2026-10-04 Jairu added, the University of Sao Paulo's 96-GPU NVIDIA B200 AI cluster inaugurated in February 2026, Latin America's largest Blackwell B200 installation

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
inforchannel.com.br	names system, part and count	NVIDIA B200 SXM, NVIDIA InfiniBand NDR (Quantum-2) · counts: 96
telesintese.com.br	names system, part and count	NVIDIA B200 SXM, NVIDIA InfiniBand NDR (Quantum-2) · counts: 96

Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.