NVIDIA GTC26 Berlin Dassault Systèmes and Vertiv’s AI Factory Plan : US Pioneer Global VC DIFCHQ SFO NYC Singapore – Riyadh Swiss Our Mind

Elisa Prisner, Executive Vice President at Dassault Systèmes, details how the 3DEXPERIENCE platform and Vertiv are validating NVIDIA DSX AI factory designs

Dassault Systèmes and Vertiv Holdings Co. have announced a collaboration that allows AI factory infrastructure to be configured, simulated and validated before a single piece of equipment is built on site.

The partnership pairs Dassault Systèmes’ 3DEXPERIENCE agentic platform with Vertiv’s critical digital infrastructure expertise, targeting faster delivery of NVIDIA DSX AI factories.

AI factories combine power, cooling, network, controls and compute into a single tightly coupled system. The virtual twins developed by the two companies spot errors before they become major.

Vertiv AI factory equipment data rendered in a virtual twin using Dassault Systèmes’ 3DEXPERIENCE platform. Credit: Vertiv & Dassault Systèmes

Virtual twins replace guesswork

A change in one domain can create unforeseen downstream effects in another, meaning operators traditionally only discover problems once construction is under way.

The collaboration addresses this through:
  • Extending Vertiv’s SmartRun configurable virtual twin across broader AI factory infrastructure, including Vertiv OneCore
  • Using the 3DEXPERIENCE platform for systems engineering, simulation and virtual twin modelling
  • Validating Vertiv’s pre-engineered SimReady power and cooling reference designs against the NVIDIA Vera Rubin DSX AI Factory reference design and NVIDIA Omniverse DSX Blueprint
  • Representing Vertiv OneCore’s integrated power, thermal management and supporting infrastructure as a single digital system
  • Giving engineers traceability over how design changes in one discipline affect performance elsewhere in the system

Scott Armul, Chief Product and Technology Officer at Vertiv, says the approach reflects the scale of the challenge facing data centre operators as AI demand grows.

Scott Armul, Chief Product and Technology Officer at Vertiv

“The winners in AI will be those who can turn computing capacity into productive capacity fastest,” he says.

“That requires the entire AI factory to work as one system, across compute, power, cooling and physical infrastructure.

“By combining Vertiv’s pre-engineered infrastructure expertise with Dassault Systèmes’ model-based capabilities, we can configure and validate that system before construction, reduce complexity in the field and help customers bring AI capacity online faster.”

Our 3DEXPERIENCE platform allows customers to explore configurations, understand the consequences of design decisions and validate the performance of an entire system virtually, before any infrastructure exists physically.

Elisa Prisner, Executive Vice President, Strategy, Industry, Marketing and Transformation at Dassault Systèmes

From fixed designs to configurable systems

Rather than locking customers into a fixed blueprint, the model-based approach lets engineers evaluate multiple configurations, propagate design changes and generate engineering deliverables from one shared environment.

Dassault Systèmes says this preserves design intent from early configuration through construction, commissioning and future technology upgrades.

Elisa Prisner, Executive Vice President, Strategy, Industry, Marketing and Transformation at Dassault Systèmes, says: “AI factories are among the most complex engineered systems being built today.

Youtube Placeholder

“That complexity demands a systems view: modeled, simulated, orchestrated and optimised in a virtual twin, before the entire factory is built.

“Together with Vertiv, we are transforming standardised infrastructure knowledge into configurable virtual twins for NVIDIA DSX AI factories.

“Our 3DEXPERIENCE platform allows customers to explore configurations, understand the consequences of design decisions and validate the performance of an entire system virtually, before any infrastructure exists physically.”

Deployment timelines under the model

Vertiv has published modelled customer outcomes based on assumptions built into its own simulations.

https://datacentremagazine.com/news/inside-dassault-systemes-and-vertivs-ai-factory-plan?shem=aimgspe,