# Is LG Building a Serious Humanoid Robot?

LG is targeting a Q1 2027 public unveiling for a next-generation bipedal humanoid robot powered by Nvidia Jetson Thor and built on the Nvidia Isaac GR00T foundation model — and on August 13, 2026, LG Corp chairman and CEO Kwang Mo Koo and Nvidia founder and CEO Jensen Huang formalized that commitment in a memorandum of understanding signed at Nvidia's headquarters in Santa Clara, California. The agreement is broader than a single robot: it spans [Physical AI](https://humanoidintel.ai/glossary/physical-ai) infrastructure, wheel-based manufacturing robots, and automotive platforms. But the bipedal humanoid is the headline, and it represents LG's clearest public commitment yet to competing in a segment currently dominated by pure-play robotics startups and vertically integrated tech giants.

The robot will run Nvidia Halos for Robotics — described in the source as the industry's first full-stack safety system for robotics — alongside Isaac GR00T for reasoning and Jetson Thor for onboard compute. Hardware components, including actuators, sensors, and batteries, are expected to be sourced internally across LG's subsidiaries: LG Electronics, LG Innotek, and LG Energy Solution.

---

## What the LG-Nvidia MOU Actually Covers

The memorandum signed August 13 is not a product launch — it is a structured execution framework across four distinct workstreams:

1. **Bipedal humanoid development** targeting a Q1 2027 public unveiling
2. **Wheel-based robot deployment** — LG CLOiD robots validated on LG Electronics' washing machine manufacturing line in Tennessee within 2026
3. **AI factory buildout** — an AI factory reference site using Nvidia DSX and Nvidia Vera Rubin planned for the first half of 2027, followed by an 80 MW LG AI Factory in Cheonan, Chungcheongnam-do, targeted for completion in the first half of 2028
4. **Automotive platform** — a next-generation AI-defined vehicle platform built on Nvidia DRIVE Hyperion

A joint Task Force comprising technical and business experts from both companies will oversee R&D, site validation, and commercialization across all four areas.

The CLOiD deployment this year is analytically important: it gives LG real-world manufacturing data before the humanoid even hits a demo stage. That sequencing — validate the simpler wheel-based platform in production first, then bring in the bipedal — is operationally sound and suggests LG's robotics team is thinking about the pipeline in terms of data flywheel construction, not just hardware showcase moments.

---

## The Stack: GR00T, Jetson Thor, and Halos

LG's humanoid will run on a now-familiar Nvidia substrate that has become something close to a default configuration for newcomers to the bipedal space. Nvidia Isaac GR00T provides the humanoid foundation model — an open reasoning model designed to accelerate [sim-to-real transfer](https://humanoidintel.ai/glossary/sim-to-real-transfer) and reduce the data requirements for teaching robots new tasks. Jetson Thor handles onboard compute. Halos for Robotics adds the safety layer.

What distinguishes LG's approach, at least on paper, is the plan to run GR00T in parallel with its own in-house Robot Foundation Model (RFM), currently under development. LG CNS PhysicalWorks — described as LG's robot data platform — will underpin an on-site physical AI data factory supporting continuous data collection, synthetic data generation, training, and verification. The stated goal is to use GR00T to accelerate near-term capability while building proprietary model assets that compound over time.

Whether LG can execute that dual-track strategy depends heavily on the talent and data infrastructure it can actually assemble — areas where purpose-built robotics companies like [Figure AI](https://humanoidintel.ai/companies/figure-ai) or [Agility Robotics](https://humanoidintel.ai/companies/agility-robotics) have multi-year head starts.

---

## Skeptical Read: What LG Has Not Said

The MOU is conspicuously light on hardware specifics. The source provides no [degrees of freedom](https://humanoidintel.ai/glossary/degrees-of-freedom) count, no actuator type breakdown, no payload targets, no locomotion benchmarks, and no indication of where the humanoid sits on the capability spectrum relative to current-generation systems. "Q1 2027 public unveiling" is a demo milestone, not a deployment commitment.

LG has been circling humanoid robotics for years. Its CLOiD brand has existed in wheel-based form for some time, and the company's prior robotics efforts have not resulted in commercially scaled deployments. The Nvidia partnership gives LG a credible AI stack and a co-marketing relationship with the most prominent infrastructure player in the space — but it does not resolve the harder questions around whole-body control, dexterous manipulation, or the organizational focus required to ship a competitive bipedal platform.

The 80 MW AI factory in Cheonan is also a significant capital commitment, though the source does not specify the total investment figure. LG positioning that facility as a reference site it can sell to "global big tech customers" as a turnkey solution suggests the robotics program is partly a Trojan horse for LG's AI infrastructure business — which may or may not align with what it takes to build a best-in-class humanoid.

---

## Industry Implications

LG's entry into the Nvidia GR00T ecosystem adds another major consumer electronics and manufacturing incumbent to a stack that is becoming increasingly crowded. The strategic logic for Nvidia is clear: every OEM that standardizes on Jetson Thor and Isaac GR00T deepens the moat around its physical AI platform and generates the heterogeneous deployment data that improves the foundation model.

For the broader humanoid field, LG's participation matters because of what it brings in components, not necessarily in software. LG Innotek's sensor capabilities and LG Energy Solution's battery expertise are genuine assets in a hardware domain where actuator and power density remain binding constraints. If LG vertically integrates those capabilities into a competitive humanoid platform, it could reduce the bill-of-materials pressure that currently compresses margins for every startup in the space.

The Q1 2027 unveiling will be the first real data point on whether LG's "One LG" integration thesis holds up in a bipedal robot under dynamic conditions.

---

## Key Takeaways

- LG and Nvidia signed an MOU on **August 13, 2026** at Nvidia's Santa Clara HQ, attended by CEO Kwang Mo Koo and Jensen Huang
- LG's next-generation bipedal humanoid targets a **Q1 2027 public unveiling** and runs on **Nvidia Jetson Thor**, **Isaac GR00T**, and **Nvidia Halos for Robotics**
- Hardware components — actuators, sensors, batteries — are planned to be sourced internally across **LG Electronics, LG Innotek, and LG Energy Solution**
- LG CLOiD wheel-based robots will be validated on LG's **Tennessee washing machine manufacturing line within 2026**, providing real production data ahead of the humanoid debut
- An **80 MW LG AI Factory in Cheonan** is targeted for completion in the **first half of 2028**
- LG is developing its own **Robot Foundation Model (RFM)** in parallel with Isaac GR00T — a dual-track bet on both licensed and proprietary AI
- No hardware specs (DOF, actuator type, payload) were disclosed in the source material; the Q1 2027 milestone is a demo, not a commercial deployment commitment

---

## Frequently Asked Questions

**When will LG's humanoid robot be unveiled?**
LG is targeting a public unveiling in the first quarter of 2027, based on the MOU signed with Nvidia on August 13, 2026. No specific event or venue has been announced.

**What AI and compute hardware powers LG's humanoid?**
The robot uses Nvidia Jetson Thor for onboard compute, Nvidia Isaac GR00T as the humanoid foundation model, and Nvidia Halos for Robotics as a full-stack safety system.

**Is LG building its own AI model for robotics?**
Yes. LG is developing an in-house Robot Foundation Model (RFM) via its LG CNS PhysicalWorks data platform, intended to operate alongside Isaac GR00T and build proprietary model capabilities over time.

**What is LG CLOiD and how does it relate to the humanoid?**
CLOiD is LG's existing wheel-based robot platform. LG plans to deploy CLOiD on its Tennessee washing machine manufacturing line in 2026 to generate real-world production data — a stepping stone that will inform and strengthen the humanoid program.

**How does this affect the competitive humanoid landscape?**
LG's entry adds a major electronics and component manufacturing incumbent to the Nvidia GR00T ecosystem. Its internal access to sensors (LG Innotek) and batteries (LG Energy Solution) could reduce component costs, but the company has not disclosed hardware specs or deployment timelines beyond the initial unveiling.