Contents
From blueprint to execution
On August 13, LG and NVIDIA signed a memorandum of understanding at NVIDIA’s headquarters in Santa Clara, California, marking what LG frames as moving from strategic blueprints to full-bore execution across robotics, AI factories and mobility ¹. LG Corp chairman and CEO Kwang Mo Koo and NVIDIA founder and CEO Jensen Huang attended the signing; the two had first met in June at the LG Twin Towers in Seoul to explore physical-AI cooperation ²³. Koo said that as discussions moved fast, LG’s tasks and goals in AI factories and physical AI had become clear, and that the group would accelerate AI adoption by building industry-leading references ¹. Huang described physical AI as endowing machines with real-world understanding, reasoning, and the capacity to act safely alongside people, and said LG’s leadership in product engineering and manufacturing was being combined with NVIDIA technology to hasten the next era of robots, AI factories and autonomous vehicles ¹. The agreement is an MOU, a statement of intent: the dates and capabilities that follow are plans, not delivered outcomes.
The humanoid: a One LG body on an NVIDIA brain
Robotics is the first visible output. LG is developing a next-generation bipedal humanoid targeted for a public unveiling in Q1 2027, built on NVIDIA Isaac GR00T, NVIDIA’s open humanoid foundation model ¹³. The robot is powered by NVIDIA Jetson Thor for onboard compute, advanced reasoning and control, and uses NVIDIA Halos for Robotics, which the company describes as the industry’s first full-stack comprehensive safety system ¹. Key hardware comes from ‘One LG’ affiliates: actuators from LG Electronics, sensors from LG Innotek and batteries from LG Energy Solution ²³. LG Electronics has set up a pilot production line at its Changwon plant and begun initial production ³. At the signing, Koo presented Huang with a miniature of the humanoid bearing their signatures ²³. Ahead of the humanoid debut, LG plans to deploy the wheeled LG CLOiD robot on a washing machine line at its Tennessee plant by year’s end for real-world validation ¹³. The pair will build a ‘physical AI data factory’ on LG CNS’s PhysicalWorks robot data platform, covering data collection, synthetic data generation, training and verification ¹². Factory-floor data will also feed LG’s in-house Robot Foundation Model (RFM), now under development ¹². A joint task force of technical and business experts will drive R&D, site validation and commercialization ¹. Separately, LG has been building toward this: it unveiled its home robot CLOiD at CES 2026, set up a CEO-led Robotics Business Center in June, and runs hundreds of CLOiD units at a robot data factory in Seoul to collect training data — context on existing operations, not a new announcement ³.
AI factories and AI-defined vehicles
Beyond robots, LG plans ‘AI factory’ reference sites pairing the NVIDIA DSX architecture with One LG data-center infrastructure capabilities — thermal management, power, design and operations ¹². A reference site powered by NVIDIA Vera Rubin is planned for the first half of 2027, and LG aims to finish an 80MW LG AI Factory in Cheonan, South Chungcheong Province, within the first half of 2028 — a build explicitly tied to advancing physical AI and robot foundation models ¹³. The Korea Herald adds that Cheonan will lean on prefabricated modular methods to shave more than 20 percent off the construction schedule ³. LG expects to offer validated, turnkey One LG packages to global big tech customers ¹. In mobility, LG plans a high-performance computing platform for next-generation AI-defined vehicles built on NVIDIA DRIVE Hyperion, integrating LG’s in-vehicle infotainment and automotive software with NVIDIA’s AI-based autonomous driving ¹. The strategy, per the Seoul Economic Daily, is to expand LG’s component business — now centered on IVI — into self-driving and supply global OEMs with next-generation solutions that combine autonomy with in-cabin AI services ². The reference-site and Cheonan timelines are plans, not completed milestones ¹.
What the alliance signals
The deal is best read as a signal in the race to industrialize robotics. Industry observers cited by the Seoul Economic Daily view the MOU as elevating LG-NVIDIA relations from technical cooperation to ‘joint business partners,’ and some assess that Koo’s One LG strategy, combined with NVIDIA’s physical AI ecosystem, lays the groundwork for LG to expand beyond home appliances and automotive electronics into an industrial infrastructure provider in the AI era ². The central method and result claims here — the GR00T-based humanoid, the Q1 2027 debut, the Tennessee and Cheonan plans, and the DRIVE Hyperion platform — come from LG’s own August 14 press release ¹, with timelines confirmed by independent Korean coverage ²³. These are forward-looking plans under an MOU; whether they materialize on schedule is not yet verified.