LG and NVIDIA Advance Humanoid Robot Collaboration, Bipedal Robot Set for Q1 2027 Launch
Nashnova编辑部
LG and Nvidia signed a strategic MOU on August 13 to co-develop a bipedal humanoid robot shipping in Q1 2027, powered by Nvidia's Jetson Thor platform. Five days later, an Nvidia executive flew to Seoul for an on-site review — signaling the deal has already moved from paper to execution.
What is this robot, and whose tech is inside?
It is a bipedal humanoid robot. Its onboard brain runs on Nvidia's Jetson Thor — a high-performance computing chip purpose-built for robotics, handling inference, decision-making, and motion control.
On the software side it integrates two Nvidia systems: Isaac GR00T — a foundation model that teaches the robot to understand commands and act — and Halos for Robotics, a safety layer that prevents uncontrolled movements.
Hardware comes from three LG subsidiaries: LG Electronics supplies actuators (joints and muscles), LG Innotek supplies sensors (eyes and touch), and LG Energy Solution supplies batteries. This means → LG has built the robot as an internal supply chain, not a patchwork of outside parts.
What is the "data factory," and why does it matter?
LG plans to open a data factory at its Yangjae R&D campus in Seoul this year — a dedicated facility that simulates home and factory environments so robots can practice movements and tasks while generating training data.
In plain terms = think of it as a driving school for robots: drill enough in simulation, then graduate to the real world. The data feeds LG's in-house Robot Foundation Model (RFM).
LG set up a dedicated robotics business unit on July 1, reporting directly to CEO Yeo Jae-cheol, with a specialist data-factory team embedded inside. This reflects a bet that data collection and model training — not just hardware — are the core competency for a robot business.
What comes before the bipedal robot?
Before the 2027 humanoid launch, LG will deploy wheeled CLOiD robots on washing-machine production lines in Tennessee this year to validate the technology on a real factory floor.
Once validated, the rollout extends to other global manufacturing sites, then eventually into homes and commercial spaces. CLOiD is positioned as an AI-capable home robot that handles laundry tasks and interacts with connected appliances via the ThinQ platform.
This means → LG is following a staged path: wheels before legs, factory before home — proving the tech in a controlled setting before attempting the harder humanoid form factor.
The partnership goes well beyond robots?
The MOU also covers AI factories: LG will combine Nvidia's Vera Rubin architecture — Nvidia's next-generation data-center chip platform — with LG group cooling, battery, design, and power solutions. A reference site comes first; an 80-megawatt expansion at Cheonan follows in H1 2028.
On smart vehicles, the two companies plan to pair Nvidia's Drive Hyperion autonomous-driving platform with LG's in-vehicle infotainment and software to co-develop a high-performance computing platform for AI-defined cars.
In plain terms = Nvidia is not just selling LG chips for one robot. It is bundling AI compute, data centers, and automotive into a single framework — tying LG's manufacturing muscle deeper into the Nvidia ecosystem.
What does the pace of this deal tell us?
The MOU was signed August 13. Five days later, on August 18, Nvidia senior director Madison Huang flew to Seoul to tour the data factory and hold closed-door talks with LG executives. Huang had already met CEO Yeo in Seoul in April to discuss "physical AI."
This reflects a partnership already in intensive execution mode, not stalled at the framework-agreement stage. The Q1 2027 bipedal-robot launch is the first hard deadline for proving whether the "Nvidia AI platform + LG hardware manufacturing" integration path can deliver.
LG's robotics footprint spans home, industrial, and commercial applications, with stakes in Robostar and Bear Robotics and plans to localize actuator production. The breadth is already laid out — the test now is depth and speed.
Content is for reference only, not financial advice.