Critical digital infrastructure company Vertiv has rolled out an energy-efficient 142kW cooling and power reference architecture for the NVIDIA GB300 NVL72 platform.
The reference architecture can be customised for bespoke data center designs to reduce planning times and risk during the buildout process.
The reference architecture for NVIDIA GB300 NVL72 is part of the Vertiv 360AI infrastructure platform, designed to help customers navigate integrated solutions for powering and cooling AI workloads and other high-performance computing applications.
Vertiv solutions are now available as SimReady 3D assets in the NVIDIA Omniverse Blueprint for AI factory design and operations.
The company said that by leveraging NVIDIA Omniverse technologies in the architecture, bridging physical and digital infrastructure, this enables real-time system-level collaboration across IT, facilities and engineering teams and allowing customers to test, optimise and validate the design before a single component is built.
The reference architecture supports rack densities up to 142kW, delivering 1.5x more AI performance, up to 50% faster on-site builds and resulting in 30% less physical space.
The architecture is also liquid cooling-ready, supporting both air- and hybrid- cooled configurations and achieving what Vertiv says is up to 70% better annual energy efficiency allowed by higher water temperatures.
“By combining NVIDIA’s advanced AI platforms with Vertiv’s expertise in power and cooling infrastructure, we’re enabling customers to deploy next-generation data centers that are more efficient, scalable, and ready for the most demanding AI workloads,” said Dion Harris, senior director of HPC and AI Infrastructure at NVIDIA.
Vertiv is currently collaborating with NVIDIA to support 800 VDC data center power infrastructure for 1 MW IT racks and beyond, with Vertiv solutions available starting in 2026.
-1.jpg&w=100&c=1&s=0)
.png&w=100&c=1&s=0)
