"In the phase of AI-based reindustrialization, the biggest opportunity lies where the information and physical worlds meet, totaling $90 trillion (about 12 quadrillion won)."Jensen Huang, Nvidia CEO
"We are living in an era of unprecedented acceleration, and this massive wave led by AI is completely changing manufacturing processes."Pascal Daloz, Dassault Systemes CEO
Nvidia and French software company Dassault Systemes have joined hands to realize physical artificial intelligence (AI), the "brain" of robots. The plan is to combine Dassault Systemes' virtual twin technology—which creates twin models identical to reality in virtual space—with Nvidia's graphics processing unit (GPU)-based accelerated computing to handle the design and operation of factories to be built around the world. The goal is to dramatically boost productivity at all factories—from semiconductor, aircraft, and automobile manufacturing facilities to AI factories essential for running AI—while eliminating errors that can arise during design and production in advance to reduce expense.
Jensen Huang, Nvidia chief executive officer (CEO), and Pascal Daloz, Dassault Systemes CEO, announced this strategic partnership at the Dassault Systemes "3DExperience World 2026" event held in Houston, Texas, from the 1st to the 4th (local time) this month. Huang said, "As reindustrialization takes place worldwide, construction of semiconductor plants, supercomputer plants, and AI factories is surging," adding, "All future factories will be built on virtual twins and high-performance computing."
Dassault Systemes presented a blueprint to implement an Industry World Model through this collaboration. It is so-called "AI that understands the real world." Unlike general-purpose AI based on large language models (LLM) such as ChatGPT, it is a model that understands physical laws and has manufacturing knowledge, making it directly applicable to real industrial sites. It is similar to the concept of physical AI that Jensen Huang earlier pointed to as the next battleground for AI.
Pascal Daloz, Dassault Systemes chief executive officer (CEO), said, "AI is now moving beyond simply predicting or generating to an era of understanding the real world," adding, "Together with Nvidia, we will build an Industry World Model that combines virtual twins and accelerated computing to help more efficiently design, simulate, and operate complex systems in life sciences, materials science, engineering, and manufacturing."
◇ Dassault Systemes teams up with Nvidia to build the "brain" of future factories
Seeing the boom in large-scale AI data center construction sparked by the AI craze as an opportunity, Nvidia plans over the next 10 years to supply GPUs for AI factories (AI-specialized data centers), secure the market for industrial AI platforms and operating systems (OS) that integrate Dassault Systemes' virtual twin technology, and take the lead. As part of that, Nvidia announced industrial AI cooperation with Germany's Siemens last month and is now seen as forming an "AI alliance" with Dassault Systemes. All are corporations strong in industrial software.
Huang said, "AI is expected to become the foundation and infrastructure of all industries, like water, electricity, and the internet," adding, "Over the next 10 years, with $85 trillion to $100 trillion invested in expanding AI infrastructure, three industries will grow explosively: semiconductor plants that make AI chips, supercomputer plants that assemble the chips, and AI factories that produce actual AI."
Huang assessed that an unprecedented expansion of industrial infrastructure led by AI is also changing the landscape of manufacturing processes. In particular, because building a single 1-gigawatt (GW) AI factory requires a massive investment approaching $50 billion (about 72.5 trillion won), the need has grown to cut unnecessary expense that can arise before and after ground is broken and to improve operational efficiency.
Huang emphasized that "before building complex and massive facilities like AI factories, if you design, simulate, verify, and operate the plant in a virtual environment in advance, you can significantly reduce expense and time," adding that AI-based virtual twins will become essential technology. The virtual twin, where Dassault Systemes has strengths, creates a "twin" of real-world objects in virtual space for simulation and uses this to predict outcomes in advance to optimize products and services.
The two companies predicted that all factory design and operations will take place within AI-based virtual twins, which will run on Nvidia's latest GPU-based accelerated computing. Huang said, "The world's largest market lies where the information and physical worlds meet," adding, "It is the first time we are directly addressing this $90 trillion (about 12 quadrillion won) market, and it is the largest among all AI opportunities," expressing high expectations.
◇ "Every engineer will have an AI companion"… rebutting the software doomsday theory
At this event, Dassault Systemes introduced the concept of a "virtual companion." It is a kind of "AI assistant" that helps the work of operators, including engineers.
Daloz said, "AI is a companion, not a black box or an autopilot," introducing the virtual companions "AURA," "LEO," and "MARIE," integrated into the company's flagship 3D (three-dimensional) CAD (computer-aided design) platform SolidWorks. When a worker enters a request in a chat window as if conversing with an AI chatbot, these virtual companions can help generate new designs based on various data such as the worker's design proposals and documents.
"AURA" helps explore ideas like popular chatbots on the market such as ChatGPT or Gemini, while "LEO" handles more specific engineering and reasoning. "MARIE" deals with scientific knowledge such as materials, chemistry, and microbiology.
Dassault Systemes repeatedly stressed that advanced AI assistants are tools to enhance engineers' capabilities, not take their jobs. It said, "AI is not a technology to replace engineers, but a 'force multiplier' that amplifies their capabilities," adding that for AI to work properly on the industrial site, engineers' creativity and judgment are essential. Manish Kumar, SolidWorks CEO, also said, "AI is just the engine; the driver who decides where to go is the engineer."
Jensen Huang also rebutted the recently rising "software doomsday theory," saying it is "completely wrong." After Anthropic added legal work based on its AI model "Claude," fear that AI would replace software and related tools swept the U.S. stock market, and the share prices of major software corporations fell one after another. He said, "Every engineer and creator will have an AI companion to help with their work, and as AI companions use various software tools, tool usage will grow exponentially."