Cadence adds NVIDIA DGX SuperPOD digital twin for AI data center

Cadence has recently enhanced its Reality Digital Twin Platform library by introducing a new digital twin model of the NVIDIA DGX SuperPOD with DGX GB200 systems. This advancement empowers data center designers to simulate next-generation AI infrastructure before making physical deployments, solidifying Cadence’s position in the rapidly expanding AI factory market. In this competitive landscape, design accuracy, power efficiency, and rapid scaling are paramount considerations.

This collaboration between Cadence and NVIDIA brings together EDA expertise and NVIDIA’s accelerated computing platforms, enabling engineers to streamline design cycles and mitigate risks in AI data center buildouts. The integration of NVIDIA’s DGX SuperPOD with GB200 systems into the Cadence Reality Digital Twin Platform is a significant milestone, catering to customers who require precise behavioral simulations of cutting-edge AI infrastructure.

Expanding the digital twin ecosystem

The Cadence Reality Digital Twin Platform is positioned as the first and only solution that models AI factories and data centers against specific service-level agreements, taking into account factors such as cost, space, energy, cooling, and environmental impact. By incorporating NVIDIA’s powerful systems, Cadence aims to provide customers with accurate simulations of high-performance AI infrastructure.

Michael Jackson, senior vice president of System Design and Analysis at Cadence, emphasized the importance of confidence in meeting design requirements for rapidly scaling AI. The addition of NVIDIA’s DGX SuperPOD with DGX GB200 systems to the platform enables designers to create behaviorally accurate simulations, ultimately reducing design time and enhancing decision-making accuracy for critical projects.

NVIDIA also underscored the significance of the collaboration, with Tim Costa, general manager of Industrial and Computational Engineering at NVIDIA, stating that creating a digital twin of the DGX SuperPOD with DGX GB200 systems is a crucial step in accelerating AI factory buildouts. This collaboration addresses the increasing pace of innovation and the shrinking time-to-service requirements in the industry.

Design and operations advantages

The Cadence Reality Digital Twin Platform allows engineers to seamlessly integrate vendor-provided models into a virtual data center that mimics its physical counterpart. This capability enables the evaluation of design trade-offs, testing of power and cooling strategies, and simulation of failure or upgrade scenarios with minimal effort. Furthermore, these models can be utilized post-deployment to monitor and optimize performance throughout the data center’s lifecycle.

In addition to supporting the NVIDIA Omniverse blueprint for AI factory design and operations, the platform now includes the NVIDIA DGX SuperPOD, further strengthening the collaboration between Cadence and NVIDIA. With a library comprising over 14,000 items from 750 vendors, Cadence offers comprehensive coverage of mainstream data center components, with the option for customers to request the creation of missing parts as part of standard support.

Showcasing at AI Infra Summit

Cadence is set to showcase the expanded platform at the upcoming AI Infra Summit in Santa Clara from September 9–11. The event will feature live demonstrations, multiple speaking sessions, including discussions on power-efficient chip design, AI for design, and simulation-driven AI data center operations. As AI factories continue to scale rapidly, the ability to validate design choices virtually before hardware implementation could be a game-changer for operators worldwide.