Dassault Systèmes and Vertiv Join Forces to Simulate AI Factories Before Construction
The editor of the 3DEXPERIENCE platform and Vertiv Holdings Co. announced on October 22, 2026 in Berlin a collaboration to configure, simulate and validate AI factory infrastructure before construction begins. The two groups aim to reduce field complexity and shorten the path from design to productive AI capacity.
3DEXPERIENCE Virtual Twins Applied to Vertiv Equipment
The collaboration combines the agentic 3DEXPERIENCE platform and Dassault Systèmes' virtual twins with Vertiv's expertise in critical digital infrastructure. It extends initial work conducted around the configurable Vertiv SmartRun virtual twin and expands it to a broader range of AI factory infrastructure solutions, including Vertiv OneCore, a hybrid data center solution integrating critical power supply, thermal management and associated infrastructure.
Specifically, Dassault Systèmes' systems engineering, simulation and virtual twin tools are used to configure and validate Vertiv's pre-established reference designs for power and cooling, dubbed SimReady. These designs are based on the Nvidia Vera Rubin DSX AI Factory reference design and the Nvidia Omniverse DSX Blueprint.
According to the press release, an AI factory is a system in which power, cooling, network, control and computing must be designed together, with changes in one domain potentially having significant effects on others. SimReady equipment thus become configurable system models, serving as starting points for customer-specific configurations. Engineers can evaluate configurations, propagate changes across disciplines, conduct system-level and multiphysics validations, and produce engineering deliverables from a common environment.
Targeted Gains Based on Vertiv Model Assumptions
The two companies are targeting objectives for customers including acceleration of up to 10 times from expressed needs to a configured and validated AI factory solution. They also aim for infrastructure deployment period reductions of up to 50%.
According to Vertiv's model assumptions, the combined approach would enable capacity to be brought into service approximately 10 weeks earlier than a conventional reference scenario modeled. This would result in up to 30% more tokens produced cumulatively within 12 months following project launch, at identical installed capacity.
The press release clarifies that this gain derives solely from the earlier availability of facilities. It reflects neither an improvement in operational efficiency nor a result measured at a customer site.
Demonstration Planned for Nvidia GTC Berlin Conference
The two groups plan to present elements of this approach at the Nvidia GTC Berlin conference, held from October 20-22. The demonstration is intended to show how a facility's requirements guide infrastructure configuration, how teams virtually evaluate power and cooling constraints, and how Vertiv uses factory-built infrastructure to accelerate physical deployment.
"The AI winners will be those who can most rapidly transform computing capacity into productive capacity," said Scott Armul, director of products and technology, cited in the press release. Elisa Prisner, executive vice president of Strategy, Industry, Marketing and Transformation at Dassault Systèmes, for her part indicated that the 3DEXPERIENCE platform enables customers to "virtually validate the performance of an entire system, before any infrastructure physically exists."
