Aug 18: LG Electronics and NVIDIA are expanding their collaboration in humanoid robotics, with a joint effort aimed at building one of the industry’s largest robot-training datasets.
LG plans to generate 100,000 hours of humanoid robot training data by the end of 2026, combining real-world operational data with synthetic and augmented datasets.
The initiative will use LG’s robotics and manufacturing expertise alongside NVIDIA’s physical AI technologies, including NVIDIA Omniverse, Cosmos and Isaac. LG is developing its Data Factory in Seoul to collect and process data from robots operating across home, manufacturing and logistics environments.
The large-scale dataset is expected to support the development of LG’s Robot Foundation Model, helping robots learn complex physical tasks and operate more effectively in real-world environments.
The collaboration marks another step in LG and NVIDIA’s efforts to accelerate the development and commercialization of next-generation humanoid robots.
