NVIDIA and Doosan Group Collaborate to Advance Physical AI and AI Factory Infrastructure
NVIDIA and Doosan Group are expanding their partnership across robotics, AI data center power, and advanced PCB materials.
Intelligence analysis by GPT-5.4 Mini
The collaboration links NVIDIA’s accelerated computing stack with Doosan’s industrial robotics, power-generation, and materials businesses. The goal is to support physical AI systems and the infrastructure needed for AI factories.
NVIDIA and Doosan are trying to build smarter robots and the big machines that power them, kind of like giving a factory a brain, muscles, and a strong battery all at once.
Analysis
What the collaboration covers
NVIDIA says it is expanding its work with Doosan Group across physical AI, robotics, and AI factory infrastructure. The partnership spans Doosan Robotics, Doosan Bobcat, Doosan Enerbility, and Doosan Corporation Electro-Materials BG, bringing NVIDIA’s accelerated computing platforms into industrial automation, power generation, and advanced electronics materials.
Robotics and physical AI
Doosan Robotics plans to use NVIDIA Isaac Sim, Isaac Lab, Cosmos open world foundation models, the Newton physics engine, and Jetson Thor in its Agentic Robot OS. The company says this platform is meant to connect perception, reasoning, simulation, learning, and on-device inference. NVIDIA and Doosan say the goal is to make industrial robots better at handling complex, changing environments and to support simulation-to-real workflows, physics calibration, and AI reasoning.
The companies also plan to explore reference use cases for tasks such as depalletizing and sanding, along with new robot forms like dual-arm and humanoid platforms. Doosan Robotics says the effort could help it move from a robot-arm maker toward a full-stack AI-first robotics company.
AI factory power and materials
Doosan Enerbility is exploring support for NVIDIA AI factories and the NVIDIA DSX platform through its power portfolio, including gas turbines, steam turbines, small modular reactors, and hydrogen fuel-cell systems. NVIDIA frames this as relevant to data centers that need reliable, high-efficiency, always-on power.
Doosan Corporation Electro-Materials BG is also part of the effort through copper clad laminate, a material used in printed circuit boards for networking gear, AI accelerators, and AI server motherboards. NVIDIA says its MGX reference architecture is designed to help system makers build modular AI factory infrastructure, and advanced PCB materials matter as performance and bandwidth rise.
Overall, the article presents the partnership as a broad industrial push: robots, power systems, and components all aimed at supporting the next wave of AI infrastructure.
Key points
- NVIDIA and Doosan Group are expanding a partnership focused on physical AI, robotics, and AI factory infrastructure.
- Doosan Robotics plans to use NVIDIA’s Isaac, Cosmos, Newton, and Jetson technologies in its Agentic Robot OS.
- The companies want to build reference uses for industrial tasks like depalletizing and sanding, plus new robot forms.
- Doosan Enerbility is exploring ways to support AI factories with power systems including turbines, SMRs, and fuel cells.
- Doosan’s materials business is part of the effort through copper clad laminate for AI servers and networking equipment.
If the collaboration works as intended, Doosan’s robots could become more adaptable in real-world factory tasks, and its industrial equipment could gain more autonomy. The infrastructure side could also help make AI data centers easier to power and build at scale.
The plans cover several difficult areas at once, from robotics software to power systems and advanced materials, so execution risk is high. The article also describes exploration and future collaboration, which means the most important outcomes are still unproven.
