Dassault And Nvidia Bring Industrial World Models To Physical AI

During his more than two decades with Nvidia, Rev Lebaredian has had a ringside seat to the show that has been the evolution of modern AI, from the introduction of the AlexNet  deep convolutional neural network that made waves by drastically lowering the error rate at the 2012 ImageNet challenge to the introduction of generative AI and now agentic AI, where systems can create AI assistance to help with knowledge work.

That said, the next step will be even more significant.

“The real value of AI is going to express itself when we apply it to the physical world, in the era that is coming that we call ‘physical AI,’” Lebaredian, vice president of Omniverse and simulation at Nvidia, said during a press briefing this week. “With physical AI grounded in the laws of physics, AI that understands the physical world and how things in the world operate, we can unlock incredible use cases from design and engineering, digital biology, material sciences, and the ultimate expression of physical AI, which is general robotics.”

However, he added, to do this, “we first have to model the world inside a computer. We need to represent the physical world accurately so that we can design, build, and operate things in the real world.”

For this, Nvidia is expanding its 25-year partnership with Dassault Systemes, a French company that over the past 45 years has developed digital twin technologies. The two companies will combine Nvidia’s AI infrastructure, open models, and software libraries with Dassault’s latest digital-twin technologies to create a shared AI platform that will become the foundation for the development of what they’re calling “industry world models.”

The companies announced the enhanced partnership at Dassault’s annual user conference – 3DExperience World – this week in Houston.

As defined by Nvidia and Dassault, these models will be able to simulate and operate highly complex systems. from tiny molecules in drugs to massive manufacturing facilities. The goal is to give industries like biology, materials science, engineering, and manufacturing scalable, accelerated, and intelligent simulation capabilities based on a unique industrial AI architecture, science-validated world models, and grounded in science and industrial knowledge.

“World models allow AI to understand cause and effect, imagine possible futures, and make smarter decisions,” Florence Hu-Aubigny, Dassault’s executive vice president of R&D, told journalists. “Industry world models go further. They embed the first principle – physics, engineering laws, and system constraints – with four decades of industrial knowledge and know-how we’ve accumulated with our clients. They combine multi-scale, multi-discipline modeling and simulation with AI, spanning materials, components, machines, factory, and the entire industrial ecosystem, including causality, reasoning, and intent. These science-validated AI models guarantee that everything they generate aligns with the real-world industrial world.”

It comes a year after Dassault introduced its 3D Universes – which the company represents as “3D UNIV+RSES” – that include its latest digital twin technologies, generative AI to automatically generate and refine designs, and real-time data from sensors and IoT devices integrated with the virtual twin.

“With 3D Universes, our new value proposition is to become the knowledge factories, factories for learning and generating,” Hu-Aubigny said. The company’s Gen7 platform – introduced last year – and “3D universes are becoming the virtual twin factories of the world, powered by industry world models, revealed by a new way of working the virtual companion to accelerate by effective and sustainable and efficient innovation.”

Dassault’s virtual companions are key to Dassault’s 3D Universe vision, AI-based and context-aware assistants for industrial users that are integrated into its 3DExperience platform. The company. The company has three of them – Aura for business, Leo for engineering challenges, and Marie, which provides deep scientific expertise – that understand intent, can reason with the industry world model, and orchestrate actions.

What Dassault needed to propel this vision of virtual twins, industry world models, and virtual companions were AI factories and technologies at scale, she said, leading to the expanded Nvidia partnership. It will be a step forward for these industries, creating simulation environments built for their businesses rather than trying to adapt more general-purpose AI to suit the job.

This comes at a time when manufacturers are speeding up their adoption and use of AI. Rockwell Automation found in its State of Smart Manufacturing report last year that 56 percent of manufacturers are piloting AI-based smart manufacturing technologies, while 20 percent are using them at scale and another 20 percent are planning to invest. In addition, 95 percent have invested or plan to invest in AI, machine learning, generative AI, or casual AI in the next five years.

With the expanded partnership, Dassault will establish and deploy AI factories built using Nvidia’s infrastructure and offered to organizations under its Outscale cloud brand on three continents. The AI factories will include the ability to run AI models on Dassault’s 3DExperience platform from the cloud to ensure the sovereignty of the users, ensuring data privacy and intellectual property protection.

Meanwhile, Nvidia will use Dassault’s model-based systems engineering (MBSE) technology to design AI factories, with its Rubin platform being up first. The MBSE technology also will be integrated into Nvidia’s Omniverse DSX Blueprint, its open reference design and simulation framework for building and running gigawatt-scale AI datacenters.

The company’s pointed to several uses cases to outline the benefits of the combined technologies. Nvidia’s BioNeMo platform will work with Dassault’s Biovia world models to advance the discovery of new molecules and materials in biology and materials research, while Dassault’s Simulia AI-based virtual twins will use Nvidia’s CUDA-X libraries and AI physics libraries to help designers and engineers more accurately and instantly predict outcomes.

Nvidia’s Omniverse physical AI libraries will be integrated into Dassault’s Delmia virtual twin for global production systems for autonomous and software-defined production systems and Nvidia’s AI technologies and Nemotron open models will be combined with the 3DExperience platform, industry world models will be combined with Dassault’s industry world models and virtual companions of the 3DExperience agentic platform.