Dassault Systèmes, Vertiv Announce Collaboration to Accelerate AI Factory Design and Buildout
TECHWORLD ·
✦ AI Summary
Dassault Systèmes announced a collaboration with Vertiv. The partnership aims to accelerate the design and buildout of AI factories by combining Dassault Systèmes' 3DExperience agentic platform with Vertiv's digital infrastructure expertise. The effort will support NVIDIA DSX AI factory infrastructure and enable simulation and validation in a virtual environment before physical buildout.
Dassault Systèmes announced that it has decided to collaborate with Vertiv. The partnership aims to accelerate the design and buildout of AI factories by combining Dassault Systèmes' 3DExperience agentic platform with Vertiv's core digital infrastructure expertise. The effort will support NVIDIA DSX AI factory infrastructure.
The two companies will support the creation of virtual environments before physical buildout, enabling simulation and validation. This approach is expected to reduce the complexity of on-site deployment by making validation possible before work begins in the field. It is also expected to shorten the time from initial design to the actual start of AI computing infrastructure operations.
Dassault Systèmes said the foundation of the collaboration is Vertiv's configurable virtual twin solution, SmartRun. It also said the companies will extend the scope of the model-based approach to a broader range of AI factory infrastructure solutions, including OneCore.
The 3DExperience platform provides model-based systems engineering (MBSE), simulation and virtual twins. The platform supports the configuration and validation of Vertiv's predesigned SimReady power and cooling reference designs based on NVIDIA Vera Rubin DSX AI factory reference design and NVIDIA Omniverse DSX blueprint.
Dassault Systèmes said an AI factory is a complex infrastructure in which power, cooling, network, control and computing systems are closely interconnected. It also said that if the design of one area changes in an AI factory, it can have a significant impact on the performance and operation of other areas.
Dassault Systèmes said that to build AI factories efficiently, companies must validate the interactions and performance of the entire system in advance, beyond individual infrastructure layers. It said this requires an integrated engineering approach.
OneCore is a data center solution based on a hybrid build method that integrates core power, thermal management and support infrastructure. The 3DExperience platform uses MBSE, simulation and virtual twin technologies to create a single integrated digital system for Vertiv infrastructure.
Dassault Systèmes emphasized that using the 3DExperience platform allows Vertiv's SimReady power and cooling assets to be transformed into system models tailored to customer requirements. It said this enables the assets to be used in design and validation.
The companies' collaboration aims to accelerate the configuration and validation process for customer requirement-based AI factory solutions by up to 10 times. It also focuses on reducing the specified infrastructure buildout schedule by up to 50%.
Vertiv assumed that applying the integrated approach could bring the start of AI computing infrastructure operations forward by about 10 weeks compared with the modeling baseline of the conventional build method. This is based on Vertiv's modeling assumptions.
The expected benefit is up to a 30% increase in cumulative AI token generation over the 12 months after project kickoff, based on the same installed capacity. Vertiv believes it can produce more AI tokens over a given period even with the same installed capacity.
Scott Armul, Vertiv's Chief Product and Technology Officer (CPTO), said the competitive edge in the AI era depends on how quickly computing capacity can be turned into actual production capability. He said that, to do this, it is necessary to configure the entire AI factory — including computing, power, cooling and physical infrastructure — as a single system.
Vertiv said that combining its predesigned infrastructure expertise with Dassault Systèmes' model-based technology makes it possible to configure and validate the entire system before physical buildout. It added that this helps reduce on-site complexity and bring forward the start of customers' AI computing infrastructure operations.
Alysa Frisner, Senior Vice President of Strategy, Industry, Marketing and Transformation at Dassault Systèmes, said AI factories are the most complex category of engineering systems being built today. She explained that, to address this complexity, an integrated approach is essential: the entire system must be modeled, simulated, orchestrated and optimized in a virtual twin environment before the physical facility is built.
Alysa Frisner also said Dassault Systèmes is working with Vertiv to turn standardized infrastructure expertise into configurable virtual twins for NVIDIA DSX AI factories. She added that the 3DExperience platform enables customers to review different infrastructure configurations, understand the impact of design decisions and validate overall system performance in a virtual environment before physical infrastructure is built.
Source: TECHWORLD · Lee Gwang-jae
Original: https://www.epnc.co.kr/news/articleView.html?idxno=408638
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.