Shop/NVIDIA GB200 NVL72 Rack System
Representative
AI Clusters and Rack-ScaleDatacentre

NVIDIA GB200 NVL72 Rack System

Rack-scale Blackwell platform with 72 GPUs and 36 Grace CPUs in a single NVLink domain.

Rack-scale
Liquid cooled
Request quote

In plain English

The established rack-scale Blackwell system. It is the practical choice where GB300 allocation is tight and the workload is large-scale training or trillion-parameter inference.

Best for

  • Large-scale training
  • Generative AI platforms
  • HPC and data processing

Full specification

You will also need

  • · Liquid cooling
  • · High-density power
  • · InfiniBand or Spectrum-X fabric

From £2,250,000

ex VAT

Alternatives worth considering

AI Clusters and Rack-Scale
Representative

NVIDIA GB300 NVL72 Rack System

An entire rack delivered as one computer. It is bought for frontier model training and very large inference platforms, and requires liquid cooling and high-capacity power.

Memory

20736 GB

Deployment

Datacentre

High-MemoryMulti-GPU

From £2,950,000

ex VAT

Allocation only

Data Centre AI Accelerators
Representative

NVIDIA B200 SXM Accelerator

The mainstream high-end training GPU. Sold as part of an eight-GPU server rather than as a card you fit yourself.

Memory

180 GB

Deployment

Rack server

DatacentreHigh-Memory

From £32,500

ex VAT

Available to order

AI Clusters and Rack-Scale
Representative

NVIDIA Vera Rubin NVL72

An entire rack that behaves as a single enormous computer. It is bought by organisations training frontier models or running very large reasoning services, and needs liquid cooling and datacentre-grade power.

Memory

20736 GB

Deployment

Datacentre

High-MemoryMulti-GPU

From £3,250,000

ex VAT

Allocation only