Vertiv releases 7MW Reference Architecture for NVIDIA Platform

Vertiv, a global provider of digital infrastructure, has announced a 7MW reference architecture for the NVIDIA GB200 NVL72 platform, developed in collaboration with NVIDIA. This architecture is designed to support the transition of traditional data centres into AI-driven infrastructures, enabling enterprises to run AI applications effectively.

Accelerated Deployment and High-Efficiency Design

The reference architecture streamlines the deployment of NVIDIA's GB200 NVL72 liquid-cooled platform, supporting power densities of up to 132kW per rack. With a comprehensive approach to infrastructure design, it focuses on optimizing deployment speed, performance, resilience, cost, energy efficiency, and scalability, meeting the needs of both current and future-generation data centres.

“We are proud to deepen our collaboration with NVIDIA to enable AI-driven data centres of today and tomorrow,” said Giordano (Gio) Albertazzi, CEO of Vertiv. “As a leader in critical power and cooling infrastructure, Vertiv is uniquely positioned to support the NVIDIA GB200 NVL72 platform. Our portfolio of high-performance power and cooling solutions, combined with our global scale, will enable customers to deploy AI data centres faster, more efficiently and with greater flexibility to address densification, dynamic workloads, retrofits and enable future-ready designs.”

“New data centres are built for accelerated computing and generative AI with architectures that are significantly more complex than those for general-purpose computing,” said Jensen Huang, founder and CEO of NVIDIA. “With Vertiv's world-class cooling and power technologies, NVIDIA can realize our vision to reinvent computing and build a new industry of AI factories that produce digital intelligence to benefit every company and industry.”

Vertiv Unveils Power and Cooling Architecture for AI Workloads with NVIDIA Blackwell Platform

Integrated Power and Cooling Solution for AI Workload Deployment
  
Vertiv, a global provider of critical digital infrastructure, has introduced a reference architecture developed with the NVIDIA Blackwell platform. This architecture enables efficient deployment of AI workloads in both new and existing data centres, reducing deployment risk and facilitating standardization. The design includes a complete critical power infrastructure, minimizing stranded power by aligning AI clusters with data centre capacity blocks. Additionally, the hybrid cooling infrastructure combines liquid and air cooling for effective high-density heat management, with options for Open Compute Project-inspired systems, such as DC power shelves.

Part of Vertiv 360AI Portfolio

This architecture forms part of the Vertiv 360AI portfolio, intended for both retrofit and greenfield data centres, with solutions that address AI and high-performance computing needs. Key features of the co-developed NVIDIA GB200 NVL72 design include:

Rapid Deployment and Retrofit: Utilizing preconfigured modules and factory integration, the Vertiv MegaMod CoolChip enables deployment of critical AI infrastructure up to 50% faster than traditional onsite builds.
  
Space-Saving Power Management: Leveraging advanced power technologies like the Vertiv Trinergy UPS and Vertiv EnergyCore lithium battery cabinet, this design optimizes space usage, reducing the required area by approximately 40% compared to legacy systems.

Energy-Efficient Cooling: Incorporating liquid and low-GWP air cooling at scale—such as Vertiv AFC chillers, Vertiv Liebert CW room cooling, and Vertiv XDU coolant distribution units—this approach can reduce annual cooling costs by up to 20%.

Dynamic Workload Management: Integrated load averaging through lithium-ion batteries and next-generation UPS supports dynamic GPU workloads.

Installation and Operations Services: With a network of approximately 4,000 field service engineers globally, Vertiv offers extensive lifecycle services for complex system support across retrofit and new builds.

Supporting AI-Ready Infrastructure

As AI adoption grows rapidly, Vertiv aims to meet the needs of accelerated computing with a comprehensive infrastructure portfolio designed for AI-specific requirements. This collaboration with NVIDIA provides a pathway for scalable deployment, setting a roadmap for future technical co-development to enable accelerated computing infrastructure at scale.

Read More:

Vertiv Partners with VMJ Lanka for Service Provider Agreement