Nvidia and Doosan Group have broadened their cooperation from robots to power and materials for AI data centres. Doosan Group said on June 8 that four affiliates, Doosan Robotics, Doosan Bobcat, Doosan Enerbility and Doosan Electronics BG, will expand cooperation with Nvidia in physical AI, robotics and AI factory infrastructure.

Doosan Group plans to expand linkages with Nvidia platforms step by step based on cooperation results at each affiliate. Doosan Robotics will integrate Nvidia's robot simulation frameworks Isaac Sim and Isaac Lab, the world foundation model Cosmos for learning real physical environments, and the edge AI computing module Jetson Thor into its Agentic Robot OS. Agentic Robot OS is an AI-based platform that links perception, reasoning, simulation, learning and on-device inference.

Doosan Robotics is pushing a shift from a robot arm manufacturer to a software-focused robot solutions company, the company said. It also plans to develop reference use cases for tasks specialised for industrial sites, such as pallet sorting and polishing work, as well as new forms such as dual-arm and humanoid robots.

Doosan Bobcat is reviewing applying Nvidia physical AI technology to construction, agriculture and logistics equipment. It aims to develop a world model that allows equipment to recognise diverse working environments and respond to changing conditions. It also set building an industry-standard ecosystem in the field of small autonomous equipment as a task to pursue.

Doosan Enerbility is reviewing ways to supply power to Nvidia's AI factory platform DSX, based on its power infrastructure portfolio that includes gas turbines, steam turbines, small modular reactors and hydrogen fuel cell systems. It plans to discuss optimising power generation facilities and assessing low-carbon power sources in response to rising power demand from AI data centres, the company said.

Doosan Electronics BG will support next-generation AI data centre infrastructure through copper-clad laminate, or CCL, a key material for printed circuit boards. High-performance CCL is used in networking equipment, AI accelerators and AI server mainboards, and requires minimal signal loss and high reliability. It aims to contribute to securing high-speed signal quality in line with the spread of Nvidia's modular accelerated system architecture MGX ecosystem.