GB300 NVL72
Rack scale, 72 liquid-cooled GPUs
| GPU | 72 Blackwell Ultra and 36 Grace (2,592 cores) |
|---|---|
| Memory | approximately 20.7 TB HBM3e (40 TB of fast memory) |
| Compute | approximately 1.1 ExaFLOPS FP4 dense |
| NVLink | 130 TB/s across a 72-GPU domain |
| Power draw | approximately 120 kW, liquid cooling |
| Form factor | full 42U rack |
What the machine runs
A single 72-chip supercomputer serves frontier models and a private AI cloud. Recommended runtime: disaggregated vLLM with LiteLLM.
Positioning
Large enterprises host their own frontier model in a single rack, in a sovereign AI cloud.
Manufacturers
Rack integration from Dell, HPE, Lenovo, Supermicro, ASUS and Gigabyte, with liquid cooling.
GPUDirect Storage
On platforms equipped with ConnectX cards, GPUDirect Storage technology establishes a direct path between NVMe or NVMe over Fabric storage and GPU memory, bypassing the CPU buffer. Sustained throughput reaches approximately 50 gigabytes per second per link. Compatible storage arrays, such as NetApp, VAST, DDN or WEKA, feed training and large-scale RAG at full speed.
Support
Open stack supported by QDNA and the communities. NVIDIA AI Enterprise option (NIM, NeMo, Triton) with a service-level agreement, per card.