LG Signs AI Cooperation MOU with NVIDIA, Plans to Expand Cheonan AI Factory to 80 Megawatts

en.Wedoany.com Reported - On August 14, South Korea's LG Group disclosed that it had signed a strategic business cooperation memorandum of understanding with NVIDIA on the 13th local time in Santa Clara, California, USA, to advance joint development and on-site validation in three areas: robotics, AI factories, and intelligent vehicles. After the two companies announced their cooperation framework in June this year, this agreement further clarified the implementation timeline for the relevant projects.

In the robotics sector, LG plans to unveil a bipedal humanoid robot in the first quarter of 2027. The robot will be powered by NVIDIA's Jetson Thor computing platform, utilizing the Isaac GR00T robot foundation model and the Halos for Robotics safety system, while integrating LG Electronics actuators, LG Innotek sensors, and LG Energy Solution batteries.

LG also plans to deploy its wheeled LG CLOiD robots to the washing machine production line at its Tennessee plant in the United States within 2026 to conduct proof-of-concept in a real manufacturing environment. The two companies will build a robotics data system covering data collection, generation, training, and validation based on LG CNS's PhysicalWorks platform.

The AI infrastructure project plans to first build a reference site based on NVIDIA's Vera Rubin platform in the first half of 2027 to validate cooling, power supply, and information technology solutions. LG intends to expand this facility to 80 megawatts in Cheonan, South Chungcheong Province, South Korea, in the first half of 2028, using prefabricated modular construction methods. LG estimates that this approach can shorten the construction period by more than 20%.

The project plans to integrate LG Electronics' data center cooling technology, LG Energy Solution's battery systems, LG CNS and LG Uplus's design and operational capabilities, as well as LS's power solutions. The completed AI factory will primarily be used for physical AI computing and robot foundation model development, and will serve as a reference facility for LG to validate integrated solutions for external data center customers.

The two companies will also develop a next-generation AI-defined vehicle high-performance computing platform based on NVIDIA's DRIVE Hyperion, extending LG's existing in-vehicle infotainment and automotive software businesses into the autonomous driving and in-car AI services sectors.