NVIDIA and Doosan Group are deepening their partnership to unlock new possibilities in physical AI, robotics, and AI factory infrastructure, involving several of Doosan's segments, including Doosan Robotics, Doosan Bobcat, Doosan Enerbility, and Doosan Corporation Electro-Materials BG.
This collaboration aims to merge NVIDIA's cutting-edge accelerated computing platforms with Doosan's expertise in industrial automation, power generation, and advanced electronics materials, effectively bolstering next-gen AI infrastructure.
Doosan Group operates at multiple levels within the AI factory ecosystem, from intelligent robotics systems to a comprehensive range of large-scale power solutions and high-tech materials for AI data centers.
The partnership will allow both entities to explore how NVIDIA’s physical AI stack, including the NVIDIA DSX AI factory platform and NVIDIA MGX, can fortify these sectors.
Enhancing Physical AI and Robotics
In the realm of robotics, Doosan Robotics is set to incorporate NVIDIA's Isaac Sim and Isaac Lab open robotics frameworks, along with the NVIDIA Cosmos open world foundation models, the open-source Newton physics engine, and NVIDIA Jetson Thor. This integration is designed to advance the Agentic Robot OS, an AI-centric platform that integrates perception, reasoning, simulation, learning, and on-device inference.
By harnessing NVIDIA’s physical AI innovations, Doosan Robotics intends to enhance its industrial robots' ability to perceive, reason, and act within complex and ever-changing environments. Techniques like simulation-to-real workflows, physics calibration, and AI-driven logic will enable collaborative robots to be more versatile and specialized for specific tasks, facilitating scalable deployment.
The two companies are also set to define reference use cases for high-value industrial applications, including depalletizing and sanding, as well as exploring innovative robot designs such as dual-arm and humanoid configurations.
Built on the Agentic Robot OS, these advancements aim to transform Doosan Robotics from a conventional robot arm supplier into a comprehensive AI-first robotics firm. This initiative aligns with Doosan Group's broader vision for physical AI, which extends its reach beyond just robotics into construction machinery and power equipment.
Doosan Bobcat is also investigating ways to implement NVIDIA's physical AI technologies in machinery for construction, landscaping, agriculture, and material handling. This initiative aims to accelerate the creation of specialized world models that empower Doosan Bobcat's equipment to adaptively recognize various operational settings, analyze shifting conditions, and perform tasks with greater autonomy. Additionally, both companies aspire to cultivate an industry-standard ecosystem for compact autonomous machinery.
Investigating AI Factory Power Solutions
On the energy front, Doosan Enerbility is looking into how its large-scale power infrastructure—featuring gas turbines, steam turbines, and small modular reactors—can complement NVIDIA AI factories and the NVIDIA DSX AI factory platform, along with Doosan Fuel Cell’s hydrogen fuel cells. These technologies are essential for powering AI data centers, which demand dependable, efficient, and uninterrupted energy supply.
Future collaborations could revolve around designing power systems suited for AI factory setups, optimizing generation technology, and assessing low-carbon energy options like small modular reactors. By aligning energy system insights with AI infrastructure needs, Doosan Enerbility aims to meet the climbing power demands arising from advanced computing.
Enhancing the NVIDIA MGX Ecosystem with Advanced PCB Materials
Doosan Corporation Electro-Materials BG is playing a crucial role in supporting the infrastructure for next-gen AI data centers through the provision of copper clad laminate (CCL), an essential material for printed circuit boards.
These high-performance CCLs are vital for PCBs found in networking devices, AI accelerators, and AI server motherboards, where minimal signal loss and exceptional reliability are paramount.
NVIDIA MGX offers a modular reference architecture for accelerated systems, enabling system manufacturers and collaborators to develop servers and rack-scale AI factory setups effectively. As the demands for performance and bandwidth in AI servers and networking systems grow, advanced PCB materials like CCL become increasingly critical for maintaining high-speed signal integrity across the data center equipment landscape.
For further insights into the NVIDIA DSX and MGX platforms, explore the latest developments.
Image credit: Doosan Group.
