Artificial intelligence is moving into a new phase inside industry, one defined less by isolated use cases and more by whether AI can be trusted to represent and operate the physical world. A newly announced long-term partnership between Dassault Systèmes and NVIDIA reflects that transition, positioning industrial AI not as an overlay to existing processes but as a foundational system of record.
The collaboration brings together Dassault Systèmes’ Virtual Twin technologies with NVIDIA’s AI infrastructure, accelerated computing and software platforms to establish what the companies describe as Industry World Models. These are science-validated digital representations of physical systems designed to support decision-making across biology, materials science, engineering and manufacturing. The aim is to create an industrial AI architecture that can be deployed at scale and relied upon in mission-critical environments.
At the centre of the partnership is the idea that AI must be grounded in physics, engineering constraints and validated industrial knowledge. Rather than relying solely on statistical inference or generative outputs, the combined platform is intended to understand how real systems behave, enabling professionals to design, simulate and operate complex environments with greater confidence.
From digital twins to world models
Dassault Systèmes has long positioned virtual twins as a way to mirror physical assets and processes digitally. The partnership with NVIDIA extends this concept into what the companies call world models, persistent, science-based representations that integrate data, simulation and AI reasoning.
These models are built and operated on the 3DEXPERIENCE platform, where agentic capabilities enable what Dassault Systèmes refers to as skilled virtual companions. These AI-driven assistants are designed to work alongside engineers, scientists and operators, drawing on deep industrial context to deliver insights that are explainable and traceable.
Pascal Daloz, chief executive of Dassault Systèmes, described this as a shift towards AI that understands the real world. In his view, grounding AI in science and validated industrial knowledge transforms it into a force multiplier for human expertise, particularly in sectors where errors carry high cost or risk.
For NVIDIA, the partnership aligns with its focus on what it terms physical AI, systems that operate within the laws of the physical world rather than abstract data domains. Jensen Huang, founder and chief executive of NVIDIA, said the collaboration combines decades of industrial leadership with NVIDIA’s AI and Omniverse platforms to change how large industries are designed and operated.
Infrastructure built for sovereign and scalable AI
A significant element of the collaboration lies in infrastructure. Dassault Systèmes, through its OUTSCALE brand, is deploying AI factories as part of a strategy focused on sustainability, data protection and sovereignty. These AI factories will use NVIDIA infrastructure across three continents, supporting the operation of AI models within the 3DEXPERIENCE platform while maintaining control over intellectual property and sensitive data.
At the same time, NVIDIA is adopting Dassault Systèmes’ model-based systems engineering to design its own AI factories, starting with the NVIDIA Rubin platform. This approach is being integrated into the NVIDIA Omniverse DSX Blueprint for large-scale AI factory deployment, creating a feedback loop between industrial design methods and AI infrastructure development.
The combined platform is intended to support a range of applications. In biology and materials science, NVIDIA BioNeMo combined with Dassault Systèmes’ BIOVIA world models is aimed at accelerating molecular and materials discovery. In engineering, AI-based physics behaviour modelling using SIMULIA and NVIDIA CUDA-X libraries is designed to improve predictive accuracy. In manufacturing, NVIDIA Omniverse libraries integrated into DELMIA virtual twins enable software-defined production systems.
Industry signals a shift in confidence
Several industrial organisations have already pointed to the potential impact of the collaboration. Bel Group said the combined computational and modelling capabilities support sustainable product formulation and packaging at scale. OMRON highlighted the role of autonomous, digitally validated production systems in managing manufacturing complexity.
In the automotive sector, Lucid emphasised how multi-physics digital twin simulations powered by NVIDIA’s physics-informed AI models could accelerate development without sacrificing accuracy. In aerospace, the National Institute for Aviation Research described how virtual companions could reduce certification effort while preserving information sovereignty.
Taken together, the partnership signals a broader change in how industrial AI is being framed. Rather than focusing on point solutions or experimental deployments, Dassault Systèmes and NVIDIA are positioning AI as a shared, validated foundation for industrial work. As AI becomes embedded deeper into physical systems, the ability to trust its outputs may prove more important than its novelty.
