Vertiv Unveils Gigawatt-Scale Reference Architectures for NVIDIA Omniverse DSX Deployment Flexibility

Introduction

In an era where artificial intelligence (AI) is revolutionizing industries, Vertiv has emerged as a front-runner in providing critical digital infrastructure solutions. Their recent announcement of gigawatt-scale reference architectures for the NVIDIA Omniverse DSX Blueprint marks a significant milestone. This initiative not only aims to reduce the Time to First Token but also enhances deployment flexibility across various AI platforms.

Overview of Gigawatt-Scale Reference Architectures

Vertiv's approach is rooted in a desire to offer unprecedented deployment methods. Moving away from conventional construction techniques, their reference architectures provide various options, including traditional stick-built methods, hybrid models, and fully prefabricated solutions. This broad optionality allows businesses to align their infrastructure with their specific operational requirements, thus fostering enhanced scalability and quicker deployment.

Vertiv utilizes the prefabricated Vertiv™ OneCore platform as a cornerstone for this initiative. This innovative platform integrates compute, power management, cooling systems, and support services into a cohesive system, achieving significant cost and time efficiencies. As Scott Armul, Vertiv's executive vice president, noted: “The AI industrial revolution requires a complete reinvention of data center infrastructure.”