Dassault Systèmes and Vertiv Connect AI-Infrastructure Design with Simulation

Dassault Systèmes and Vertiv have announced collaboration on configuring and assessing infrastructure for AI computing facilities before construction. The approach combines 3DEXPERIENCE modelling and simulation with Vertiv’s power and cooling knowledge, including configurable virtual representations of infrastructure. The announcement presents projected deployment benefits under a defined model. It explicitly distinguishes those projections from measured customer results and from improvements in operating efficiency. For engineers, the central issue is the coordination of several systems that must work together at facility scale.
Infrastructure changes propagate between disciplines
Compute equipment, electrical supply, cooling and physical layout impose requirements on one another. A change in the expected equipment configuration can affect available space and heat removal, while a different infrastructure arrangement can constrain future expansion. A common system model can make those relationships easier to investigate before field work fixes the interfaces in place.
The model should identify which assumptions are agreed and which remain provisional. A configuration based on an early equipment requirement may need substantial review when that requirement changes. Keeping the dependencies visible helps teams understand the consequence of a revision rather than updating individual discipline models independently and discovering the mismatch during construction or commissioning.
Configurable reference designs need valid combinations
Using pre-engineered assets can reduce the effort required to develop a facility concept. However, a configurable library needs rules describing which combinations are appropriate. A component that is valid in one arrangement may require a different connection, support or operating condition in another. The model should carry enough information to prevent an attractive configuration from concealing an incompatible interface.
Engineering review should also establish what has already been validated and what remains project-specific. A reference design can provide a useful starting point without proving the performance of every customer configuration. Site conditions, the selected load and the final control arrangement may require further analysis. The virtual model should make that remaining work easier to identify rather than implying it has disappeared.
Simulation supports earlier trade-offs
System-level and multiphysics assessments can help compare options before construction begins. Their usefulness depends on whether the analysis represents the conditions relevant to capacity, stability and service access. A layout that meets a nominal thermal target may still be difficult to maintain or expand. The evaluation should include the operational questions that influence the facility’s practical value.
The source’s projected output benefit comes from bringing capacity online earlier under its modelled assumptions. That is different from producing more computation per unit of energy once the facility is operating. Keeping those claims separate is important for investment decisions. Engineers should ask which schedule activities are assumed to change and whether the same constraints apply to their own project.
Preserve the model through commissioning
A virtual configuration becomes more useful when the installed state and commissioning evidence remain connected to it. Changes made during procurement or construction should be reflected in the model where they affect future decisions. Otherwise, a carefully assessed early representation can become a historical illustration rather than a dependable engineering reference.
The Dassault Systèmes–Vertiv collaboration is relevant because it treats AI infrastructure as a coupled system rather than a set of independent equipment purchases. Its value will be demonstrated by project-specific evidence that configuration and simulation reduce unresolved interfaces before field installation. The goal is a clearer route from requirements to an operational facility, with the basis of every projected benefit and technical decision available for review.
Source — dailycadcam.com. Source publication date 2026-10-22. Engineering analysis: The CAD Verse.
Report an image





MEMBERS · ONE ENGINEERING COMMUNITY
Engineering videos and tutorials
Engineering marketplace and professional opportunities
Courses, tutorials and learning paths
Engineering files and CAD models
Professional engineering projects
Discussion
0 CommentsShare your thoughts and engineering insights with the community.
Comments are closed for this article.
No comments yet. Be the first to share your thoughts.