LG Electronics and Nvidia are moving full speed ahead on building a “data virtuous cycle” system for robot learning. LG Electronics plans to deploy hundreds of robots at its Yangjae Data Factory in Seoul by year-end to secure 100,000 hours of training data—equivalent to roughly 12 years’ worth.
LG Electronics announced on the 18th that it met with Nvidia officials at LG Data Factory in Seocho-gu, Seoul, to review the direction of robotics collaboration and discuss commercialization plans. Attendees included Madison Huang, Nvidia’s senior director of Omniverse and Robotics product marketing; Jung So-young, CEO of Nvidia Korea; Ryu Jae-cheol, CEO of LG Electronics; Hyun Shin-kyun, CEO of LG CNS; and Jung Soo-heon, CEO of LG Science Park.
The meeting came just four days after LG and Nvidia signed a memorandum of understanding (MOU) for future strategic business cooperation at Nvidia’s U.S. headquarters on the 13th. The top-level executive agreement has now been translated into concrete collaboration plans at the actual robot development site.
Huang quietly entered the campus at around 8:40 a.m. without any public schedule, followed by Jung’s arrival about 10 minutes later. The meeting ran approximately two and a half hours until 11:12 a.m., exceeding the initially expected two hours.
The two companies reviewed the operational status of the Data Factory, which is being built with a target of full operation by year-end. The facility spans four floors—from basement level 1 to the third floor—with a total floor area of 10,000 square meters. It comprises spaces where robots learn various tasks and where collected data is verified and refined.
At the Data Factory, LG Electronics’ self-developed humanoid robot “LG CLOiD” is currently performing data generation, collection, and learning. CLOiD cleans in home-like environments, and in a manufacturing setting replicating LG’s Tennessee washing machine plant in the U.S., it moves, stacks, and assembles parts. Related data is also being collected from LG CNS’s logistics automation solutions and LG Innotek’s robot hand learning spaces.
Combining LG’s Manufacturing Data with Nvidia’s Physical AI
LG Electronics’ vision is to combine decades of manufacturing and logistics field data and skilled-worker know-how with Nvidia’s robotics technology. Using Nvidia’s Omniverse library, Cosmos open-world model, and Isaac open robotics development platform, the companies will augment and synthesize real-world data and scale it into large-scale training datasets.
The total data that LG Electronics will directly collect by year-end, combined with augmented and synthesized virtual data, amounts to 100,000 hours—approximately 12 years’ worth. The plan is to build a virtuous cycle in which high-quality data training and robot performance improvement reinforce each other, advancing the Robot Foundation Model (RFM) that governs a robot’s perception, judgment, and action.
CategoryDetailsData Factory ScaleBasement level 1 to 3rd floor, total floor area 10,000 m²Robots DeployedLG CLOiD, expanding to hundreds by year-endYear-End Training DataTotal 100,000 hours (approx. 12 years’ worth)Nvidia Technologies UsedOmniverse library, Cosmos open-world model, Isaac open robotics development platform
During his tour of the Data Factory, Huang personally wrote “AMAZING LG” on LG CLOiD. After the meeting concluded, he exited the building with Ryu and other LG executives, exchanged farewells, and got into his vehicle. As the car passed reporters, he rolled down the window and, when asked how the meeting went, smiled brightly and replied, “Incredible.” He then gave a thumbs-up and waved, while LG executives and employees remained outside the building, waving back until the vehicle departed.
Leaping Toward Becoming a Robotics Total Solution Provider
This collaboration will be driven by the Robotics Business Center, a CEO-direct organization that LG Electronics newly established last month. LG Electronics aims to expand its robot product lineup from industrial and commercial applications to home use, and to combine core components such as actuators with data capabilities to become a robotics total solution provider spanning both hardware and software.
Ryu Jae-cheol, CEO of LG Electronics, said, “Through ‘One LG’ synergy that consolidates the group’s core capabilities and strategic collaboration with global partners, we will secure physical AI competitiveness and transform into a robotics total solution provider.”
The Data Factory collaboration is also expected to create synergy with the next-generation humanoid development previously agreed upon by Koo Kwang-mo, chairman of LG Group, and Jensen Huang, CEO of Nvidia. The two met at Nvidia’s U.S. headquarters on the 13th and signed an MOU to develop a next-generation bipedal humanoid next year, with a target unveiling in the first quarter of next year. The two companies also agreed to deploy CLOiD at LG Electronics’ Tennessee plant in the U.S. for validation by year-end and to collaborate on building an AI factory next year.