Accelerate AI reasoning, training, and inference with NVIDIA GB300 NVL72, a rack-scale AI platform built on NVIDIA Blackwell Ultra architecture for next-generation AI factories and trillion-parameter models.
.png)
User Responsiveness
Up to 10X higher TPS per user compared to Hopper-based systems.
Unified NVLink Architecture
Connected through NVIDIA NVLink into a single rack-scale AI domain.
AI Factory Output
Delivers up to 50X greater AI factory output compared to Hopper-based system
Unified Memory Architecture
Built to support trillion-parameter AI models and large-scale reasoning workloads.
GB300 NVL72 combines next-generation NVIDIA Blackwell Ultra GPUs, high-bandwidth NVLink connectivity, and rack-scale AI infrastructure to deliver exceptional performance for AI factories, reasoning models, and large-scale enterprise AI workloads.

