[Edaily Reporter Kim Hyun-ah ] NC AI, a company specializing in physical AI, has formed a consortium with LGELECTRONICS(066570)to develop an autonomous humanoid robot based on domestically produced on-device AI semiconductors.

#NC AI, a subsidiary of NC, announced on the 3rd that it had been selected as the implementing agency for the humanoid category of the “K-On-Device AI Semiconductor Technology Development Project,” led by the Ministry of Trade, Industry and Energy and the Korea Institute for Advancement of Technology (KEIT).

This project aims to develop autonomous humanoid robot technology capable of recognizing the surrounding environment and performing human-level tasks in complex everyday spaces using domestically produced on-device AI semiconductors.

NC AI will be responsible for building the initial training dataset for learning-based GMT (General Motion Tracking) and establishing a “Human-to-Robot Retargeting” pipeline to transfer human movements to robots.

Retargeting is a technology that converts human movements to match a robot’s body structure and joint characteristics, thereby enabling the robot to perform those actions. Since the body proportions and joint ranges of motion differ between humans and robots, physical conditions—such as foot support and balance—must also be taken into account.

Recently, the robotics industry has been evolving beyond performing repetitive tasks in controlled environments toward a focus on physical AI, where robots perceive their surroundings and make autonomous decisions and movements based on the situation. Consequently, the importance of not only robot hardware but also AI models, training data, and the technology that translates these into actual movements is growing.

NC AI plans to leverage the motion data and animation production technology it has accumulated to convert a wide range of human work motions into data usable for humanoid training. The goal is to enhance the full-body motion control performance of humanoids through this process.

The company will also build a large-scale motion dataset by virtually simulating various work scenarios in a digital twin environment. Subsequently, by integrating its proprietary world model technology, the company plans to advance the technology so that robots can predict changes in their surroundings and continuously perform complex, multi-step tasks.

For this project, consortium members—including LGELECTRONICS—will collaborate to integrate robot hardware, AI models, training data, and domestically developed on-device AI semiconductors into a single development framework. On-device AI performs AI computations within the robot itself without sending data to external servers, making it well-suited for humanoid robots where real-time control and low-power operation are critical.

The developed technology is expected to be applied to general-purpose humanoid robots capable of performing various tasks—such as organizing and transporting objects, serving, and manipulating tools—in non-standard environments like homes. Its scope of application could expand to fields requiring a high level of judgment and autonomous control, including disaster response, facility inspections, and work in hazardous areas, in addition to task support in manufacturing and logistics settings.

NC AI is expanding its Physical AI business with a focus on developing world models and robot foundation models. The company plans to refine AI models capable of adapting to diverse environments and strengthen its technological capabilities to enable robots to make independent judgments and translate them into action.

Lee Yeon-soo, CEO of NC AI, stated, “Participating in this project is highly significant because the motion control technology and data-building capabilities we have accumulated over many years are being realized through the autonomous movements of actual humanoids.” She added, “We will collaborate with partners such as LGELECTRONICS to enhance the competitiveness of humanoids based on domestically developed on-device AI semiconductors and contribute to securing national strategic technologies.”