Correction: Rumored Giant Cluster Is 100K GPUs, Not 100K NVL72 Racks

A widely circulated claim about a massive training cluster was corrected: it actually involves 100,000+ GPUs across roughly 1,389 NVL72 racks, not 100,000 NVL72 servers (which would mean 7.2 million Blackwell GPUs).

2026-09-07 ~ 2026-09-07 · 2 related posts

• It's 100,000 GPUs, not NVL72 racks: viral cluster-size claim corrected — firstadopter · 2026-09-07
• Correction: 100K+ GPUs across NVL72 racks equals about 1,389 racks, not 100K racks — YuXiang_IRVL · 2026-09-07