Blackwell · 192 GB HBM3e · 700W TDP · SXM5
Blackwell
192 GB HBM3e
8.0 TB/s
3.5 PFLOPS
5th Gen 1.8 TB/s
700 W
SXM5 (HGX)
NVLink + PCIe
IN STOCK
Quote
NVIDIA B100 Tensor Core GPU — Blackwell architecture with 192 GB HBM3e memory and 8.0 TB/s memory bandwidth. The HGX B100 delivers up to 3.5 PFLOPS of FP4 tensor performance for LLM training and inference. Up to 700W TDP, NVLink 5th-gen interconnect for massive multi-GPU scaling. Drop-in for H100-class deployments with significantly higher throughput at FP4 precision.
Need NVIDIA B100 Tensor Core GPU?
NVIDIA B100 in stock. 192GB HBM3e for large-model training and inference. Volume pricing available.
Request Quote