Nvidia is accelerating its physical AI strategy in South Korea through a high-profile partnership with LG Group, signaled by Madison Huang's visit to LG's robotics data hub in Seoul just four days after her father, CEO Jensen Huang, signed a strategic agreement with LG's chairman. The rapid deployment of executives to Korea underscores how seriously Nvidia is treating the race to dominate robot training data, a critical resource for building smarter physical AI systems.

What Is the LG Data Factory and Why Does It Matter?

Madison Huang, senior director of Omniverse and robotics product marketing at Nvidia, visited the LG Data Factory at LG Electronics' Yangjae R&D Campus in Seoul on Tuesday morning, just four days after LG Group Chairman Koo Kwang-mo and Jensen Huang signed a memorandum of understanding (MOU) at Nvidia's headquarters in Santa Clara, California, on August 13. The timing signals that the two companies are moving with unusual speed to translate their partnership into concrete robotics collaboration.

The LG Data Factory is the nerve center of LG Group's robotics research and development. The facility spans one below-ground floor and three above-ground floors, covering a total of 10,000 square meters. Currently, about 50 units of LG Electronics' humanoid robot CLOiD are deployed there, performing cleaning tasks in spaces designed to replicate real homes and carrying out parts-handling, stacking, and assembly training in manufacturing environments that mirror LG's washing machine factory in Tennessee.

LG Electronics plans to increase the number of robots deployed at the facility to several hundred by year-end to generate higher-quality training data. The company aims to reach a total volume of 100,000 hours of training data by year-end, combining data collected directly at the Yangjae Data Factory with virtual data generated through amplification and synthesis. To put that in perspective, 100,000 hours is roughly equivalent to 12 years' worth of continuous data.

How Are Nvidia and LG Building a Robot Training Data Flywheel?

• Manufacturing Data Integration: LG Electronics combines decades of high-quality data accumulated from manufacturing and logistics sites and home appliances worldwide with Nvidia's robotics solutions to synthesize and amplify that data into a self-reinforcing learning cycle.
• Nvidia Technology Stack: The partnership draws on three Nvidia tools: the Omniverse Library for simulation, the Cosmos open-world model for generating synthetic training scenarios, and the Isaac open robotics development platform for robot control and learning.
• Robot Foundation Model Advancement: Through this effort, LG Electronics plans to advance its robot foundation model, which underpins humanoid robot performance, and continuously strengthen its robotics competitiveness in real-world deployment scenarios.

The strategic focus on data quality and volume reflects a fundamental truth in AI: robots learn from examples, and the more diverse, high-quality examples they see, the better they perform in unpredictable real-world environments. LG's decades of manufacturing expertise provides exactly the kind of structured, domain-specific data that Nvidia's robotics platform can leverage.

What Does This Partnership Signal About the Physical AI Race?

The August 13 MOU between Jensen Huang and LG Chairman Koo Kwang-mo pledged cooperation across three strategic business areas: physical AI, AI infrastructure, and mobility. The rapid follow-up visit by Madison Huang to Seoul demonstrates that this is not a ceremonial agreement but an active, accelerating collaboration.

"We will secure physical AI competitiveness and reinvent ourselves as a total robotics solution provider through synergies grounded in 'One LG' which unites the group's core capabilities and through strategic collaboration with global partners," said Ryu Jae-cheol, LG Electronics CEO.

Ryu Jae-cheol, CEO at LG Electronics

LG Electronics mentioned its collaboration with Nvidia for the first time in its semiannual report this year, describing it as a "joint project aimed at combining our robot hardware technology and manufacturing capabilities with Nvidia's physical AI technology stack, advancing manufacturing-site robots from proof of concept through to real-world deployment". This public commitment signals that both companies view the partnership as strategically significant enough to highlight to investors.

The choice of the Yangjae Data Factory as the meeting venue for Madison Huang and LG executives carries particular weight. The facility is still under construction, with plans to reach full operation by year-end. By inviting Nvidia's robotics leadership to tour the facility during its build-out phase, LG is signaling transparency and commitment to the partnership while also demonstrating the scale of its robotics ambitions.

For Nvidia, the partnership represents a critical piece of its physical AI strategy. While the company dominates AI chip manufacturing and software platforms, building world-class robot training datasets requires partnerships with companies that have real manufacturing expertise and access to diverse operational environments. LG's global manufacturing footprint and robotics hardware capabilities make it an ideal partner for accelerating Nvidia's push into the physical AI market.

The robotics data flywheel that LG and Nvidia are building in Seoul could become a template for how the two companies scale physical AI globally. By year-end, the facility will have generated the equivalent of 12 years of continuous robot training data, a resource that could significantly accelerate the development of more capable humanoid robots and manufacturing automation systems.