NVIDIA and Doosan Group are expanding cooperation across physical AI and AI data center infrastructure. The direction is to strengthen AI infrastructure construction capabilities by combining robotics, power systems, and electronic materials technologies.

NVIDIA announced on the 9th that it is expanding cooperation with Doosan Group in the fields of physical AI, robotics, and AI factory infrastructure.

Major affiliates including Doosan Robotics, Doosan Bobcat, Doosan Enerbility, and Doosan Electronics BG are participating in this collaboration. It is structured to combine NVIDIA's AI platform with Doosan's industrial automation technology.

Doosan Robotics is developing an 'Agentic Robot OS' utilizing NVIDIA's Isaac Sim, Isaac Lab, Cosmos models, and Jetson Thor. This system is a robot platform that integrates perception, reasoning, simulation, learning, and on-device reasoning.

Doosan Bobcat is considering applying physical AI technology to construction, agricultural, and logistics equipment. Through this, we are promoting the development of a system in which equipment recognizes the environment and performs tasks autonomously.

The two companies also plan to jointly develop robot application cases applicable to industrial processes, such as logistics loading and surface processing.

Doosan Enerbility is participating in collaborations in the field of power infrastructure for AI data centers. It provides stable power supply technology based on gas turbines, steam turbines, Small Modular Reactors (SMRs), and hydrogen fuel cells.

As AI data centers require a continuous power supply for high-performance computing, power generation facilities and power efficiency improvement technologies are cited as important factors.

Future cooperation is expected to expand to include power supply design, optimization of power generation facilities, and review of the application of low-carbon power sources.

Doosan Electronics BG participates in AI data center infrastructure through Copper Clad Laminate (CCL), a core material for printed circuit boards (PCBs).

High-performance CCL plays a role in ensuring signal integrity and stability in AI servers, network equipment, accelerator systems, and the like.

NVIDIA MGX is a modular architecture for building server and rack-level AI infrastructure, and Doosan takes on the role of materials and infrastructure in that ecosystem.