At NVIDIA GTC 2026, while NVIDIA CEO Jensen Huang was introducing the NVIDIA Vera Rubin DSX reference design system on stage, he highlighted the use of technologies from a number of partners, including Dassault Systèmes, Cadence, Siemens, and PTC. During the week of the conference, NVIDIA also announced partnerships with Cadence, Dassault Systèmes, Siemens, and Synopsys, among others. To get the details on these partnerships and what outcomes to expect, we speak to Tim Costa, GM and VP of i...
At NVIDIA GTC 2026, while NVIDIA CEO Jensen Huang was introducing the NVIDIA Vera Rubin DSX reference design system on stage, he highlighted the use of technologies from a number of partners, including Dassault Systèmes, Cadence, Siemens, and PTC. During the week of the conference, NVIDIA also announced partnerships with Cadence, Dassault Systèmes, Siemens, and Synopsys, among others. To get the details on these partnerships and what outcomes to expect, we speak to Tim Costa, GM and VP of industrial and computational engineering at NVIDIA.
The four simulation software titans all share “an enthusiasm for integrating accelerated computing” in their solutions, Costa said. Their integration “isn’t just about improving the total cost of ownership but about transforming what’s possible with real-time simulation.”
Costa also pointed out that agentic AI, made possible with accelerated computing, can “compress the product development lifecycle and improve time to market.” Perhaps the most important factor is the possibility to deliver real-time digital twins. “If your physics simulation takes three days, for example, that won’t be real time. You need to get to the point where you can solve your fluid, structural, acoustic, and thermal problems in real time,” he reasoned.
In immersive simulation exercises on NVIDIA Omniverse, the simulation software makers’ physics solvers, and NVIDIA’s AI-physics framework work in concert to enable real-time digital twins, Costa explained. “Companies that work in CAE (Computer-Aided Engineering) and EDA (Electronics Design Automation) can integrate our core CUDA-X libraries. That’s true of our libraries for first principle solver acceleration … hundreds of libraries like cuDSS [GPU-accelerated Direct Sparse Solver] and cuEST [for Electronic Structure Theory] .. We’re really just one click below [the simulation vendors’] solvers, working to accelerate their workloads and connect the real and the digital world,” he said.
At GTC San Jose, NVIDIA also unveiled the Vera Rubin DSX AI Factory platform, which includes DSX Sim to validate AI factories as High-Fidelity Digital Twins. ISV partners -- Cadence, Dassault Systèmes, Siemens, and Synopsy -- can integrate their own simulations and models for designers and builders of DSX AI factories.
“There’s tremendous hunger in the design and engineering space to adopt AI-- agentic AI, physical AI, and AI physics,” Costa said.
For his full comments, listen to the podcast.
View more
