NVIDIA & Dassault Systèmes: Accelerating Industrial AI with Digital Twins & Omniverse

The convergence of industrial AI, digital twins, and accelerated computing is reshaping how companies design, simulate, and optimize products and processes. Recent advancements, particularly around NVIDIA Omniverse and OpenUSD, are accelerating this shift, allowing for more accurate and efficient virtual environments for testing and innovation. This isn’t simply about faster design cycles; it’s about creating a feedback loop between the physical and digital worlds, enabling continuous improvement and the development of more sustainable solutions.

Bridging the Physical and Digital with Virtual Twins

At the heart of this transformation are digital twins – virtual replicas of real-world environments, facilities, and processes. These aren’t merely visual representations; they’re dynamic simulations capable of mirroring the behavior of their physical counterparts. NVIDIA, in partnership with Dassault Systèmes, is pushing the boundaries of digital twin technology by integrating Dassault Systèmes’ Virtual Twin platforms with NVIDIA’s accelerated computing, AI physics open models, and Omniverse libraries. This collaboration aims to empower designers and engineers to innovate faster and deliver more sustainable products.

Dassault Systèmes’ SIMULIA software now leverages NVIDIA CUDA-X and AI physics libraries to provide AI-based physics behavior within virtual twins. This means designers can accurately and instantly predict outcomes in simulation, reducing the demand for costly and time-consuming physical prototypes. The integration extends beyond software; NVIDIA is adopting Dassault Systèmes’ model-based systems engineering technologies to accelerate the design and deployment of large-scale AI factories, while Dassault Systèmes will deploy NVIDIA-powered AI factories on its OUTSCALE sovereign cloud, addressing data residency and security concerns.

Applications Across Diverse Industries

The impact of these advancements is already being felt across a range of industries. Lucid Motors is utilizing this technology to accelerate innovation in electric vehicles, combining simulation, AI physics models, and digital twin technology with Dassault Systèmes’ tools for vehicle and powertrain engineering. In the life sciences, researchers are employing virtual twins and NVIDIA’s BioNeMo platform to speed up molecule and materials discovery, as well as therapeutics design. The Bel Group, a major cheese producer, is using these technologies to develop healthier and more sustainable food options, specifically exploring non-dairy protein alternatives. They are using industry world models to study food proteins, accelerating research and development.

The benefits aren’t limited to product development. Omron is leveraging virtual twins and physical AI to design and deploy automation technology with greater confidence, streamlining the shift toward digitally validated production. Even in aerospace, researchers at Wichita State University’s National Institute for Aviation Research are using virtual twins and AI companions to accelerate the design, testing, and certification of aircraft. These examples demonstrate the broad applicability of digital twin technology and its potential to revolutionize various sectors.

The Power of Physics-Based World Models

A key component driving these advancements is the development of physics-based Industry World Models. These models, developed by Dassault Systèmes, are trained with extensive knowledge in fields like biology, physics, and material sciences. This allows them to accurately simulate real-world environments and scenarios, enabling teams to test industrial operations end-to-end – from supply chains to retail shelves – before implementing changes in the physical world. These models can support workflows ranging from DNA sequencing to strengthening materials used in vehicle manufacturing.

As Pascal Daloz, CEO of Dassault Systèmes, stated during his 3DEXPERIENCE World keynote, “Knowledge is encoded in the living world. With our virtual twins, we are learning from life and are also understanding it in order to replicate it and scale it.” This highlights the ambition to not just simulate reality, but to learn from it and apply those learnings to create more efficient and sustainable solutions.

OpenUSD and the Future of Simulation

Underpinning much of this progress is OpenUSD (Universal Scene Description), a powerful open standard for describing and connecting complex 3D worlds. NVIDIA’s investment in OpenUSD, alongside improvements in rendering, neural reconstruction, and world foundation models, is accelerating the construction of digital twins at scale. The latest Omniverse software development kits bridge MuJoCo and OpenUSD, allowing developers to simulate robots across platforms. Omniverse NuRec libraries and AI models enable ray-traced 3D Gaussian splatting, allowing for the capture, reconstruction, and simulation of the real world using sensor data. NVIDIA Isaac Sim 5.0 and Isaac Lab 2.2, open-source robot simulation and learning frameworks, are now available on GitHub.

Expanding Capabilities with NVIDIA GTC

For those interested in learning more, NVIDIA GTC, running March 16-19 in San Jose, offers a deep dive into industrial and physical AI. The conference will feature a keynote address from NVIDIA founder and CEO Jensen Huang, as well as an agenda packed with hands-on sessions, customer stories, and live demos. Attendees can explore a dedicated industrial AI agenda, delve into OpenUSD with a session focused on physical AI simulation, and learn from Dassault Systèmes’ Florence Hu-Aubigny on how virtual twins are shaping the next industrial revolution. A developer community livestream will also be available, allowing participants to ask questions and interact directly with NVIDIA engineers.

The integration of virtual and physical worlds, driven by technologies like NVIDIA Omniverse, OpenUSD, and advanced AI models, represents a significant leap forward in industrial innovation. As these technologies mature and develop into more accessible, we can expect to see even more widespread adoption and transformative applications across a diverse range of industries. The ongoing development and refinement of these tools will be crucial for unlocking their full potential and realizing a future where design, simulation, and optimization are seamlessly integrated.