AI summary
Summary of article:
Facing the rise of humanoid robots in China, LG and Samsung respectively used data and platform, hardware and factory scenarios to gain a foothold in the future humanoid robot market.
face chinahumanoid robot Growing rapidly, focusing on sports performance and value advantage,south korea SAMSUNG Electronics and LG Electronics also pursue differentiation strategies. Among them, LG focuses on acquiring robot training data and creating integrated solutions through cooperation with global technology companies; Samsung focuses on precision hardware and control technology and tests its commercial feasibility in its own production plants. Efforts are being made to take advantage of the opportunity in the possible future development track of humanoid robots.
Analysts believe that this competition is gradually turning into a comprehensive game focused on data, platforms, real-scenario task execution capabilities, and mass production capabilities.
South Korea’s “BusinessKorea” reported that LG Electronics is tapping into a storehouse of training data to optimize the performance of large artificial intelligence models that control robot decision-making and actions. In its earnings conference call on October 28, LG revealed that a 10,000-square-meter “data factory” located in Seoul’s Yangjae R&D Park will be fully operational in the fourth quarter. The facility will be dedicated to collecting data and training robots to complete various tasks in real environments.
LG aims to accumulate 500,000 hours of robot training data this year, five times more than the original target of 100,000 hours. The company plans to combine data collected from real environments with virtual data to improve the robot’s understanding and ability to act in different scenarios.
LG is also expanding its cooperation with leading AI company Huida. In August last year, Madison Huang, senior director of physical AI platform product and technology marketing of NVIDIA, visited the LG data factory and the two parties discussed potential cooperation directions. LG plans to launch a bipedal humanoid robot based on Huida’s Isaac GR00T robot AI model in the first quarter of 2027.
Use HUDA Omniverse, Cosmos, and other platforms to complete robot training in virtual environments and hone the capabilities needed for real scenarios.
LG is also promoting hardware research and development in parallel. In the first half of this year, the company built a test production line for actuators, the main component of robot joints, at its Changwon factory. LG Innotek provides camera and sensor technology, LG New Energy provides output experience in the battery field, and the capabilities of the group’s various business areas are being integrated.
The LG AI Research Institute is also developing “physical AI” technology to allow robots to sense the surrounding environment and take action accordingly.
Kim Chang-tae, Chief Financial Officer of LG Electronics, has clarified the company’s transformation strategy, which is no longer limited to just an equipment supplier, and is transforming into a comprehensive solution provider covering core components, hardware, software platforms and data services.
On the other hand, Samsung Electronics is focusing on precision control technology and robotic hardware iteration in industrial scenarios. On October 27, on the fourth anniversary of Samsung Electronics Chairman Lee Jae-yong’s inauguration, he visited the Robot Experience (RX) business development office in Seoul’s Ummyon-dong R&D Park to observe the progress of the development and commercialization of humanoid robots.
The RX Business Promotion Office was established in July this year under the direct supervision of Lu Taiwen, head of the Device Experience (DX) division. It integrates robot-related capabilities that were previously dispersed within the company and coordinates long-term strategy, technology research and development, and commercialization. This is also Lee Jae-yong’s first visit since the establishment of the department.
One of Samsung’s main research and development directions is to simulate the “tendon-type precise control mechanism” of human tendons to improve the movement and force control capabilities of the robot’s fingers, so that it can accurately grasp and manipulate parts of various sizes and shapes, and ultimately create an industrial humanoid robot that can complete repetitive operations and accurately handle various parts in the factory workshop.
Samsung plans to use its own production plants as the first testbed for robotics technology. It will first deploy robots in repetitive or high-risk processes in semiconductor and home appliance factories, and then use operational data collected on site to continuously optimize performance. Chris Hauser, director of the Robotics Laboratory under the RX Business Promotion Office, also proposed a commercialization path: first explore the actual implementation of industrial robots in manufacturing scenarios.
Samsung aims to transform all production plants around the world into AI-powered independent chemical plants by 2030. After completing the verification of reliability and production efficiency in a manufacturing environment, the company will prioritize the B2B market and then expand into the field of home and service robots.
Summary FAQ
-
LG believes that the core of humanoid robots lies in training data and decision-making capabilities, so it collects real scene data through data factories and combines it with the HUDA platform and virtual training to improve the robot’s ability to understand the environment and complete tasks.
-
Samsung focuses on precision hardware and control technology, especially the precision control mechanism that mimics human tendons, allowing the robot to accurately grasp parts. Initially, it will be tested in repetitive, high-risk processes such as semiconductor and home appliance factories.
-
LG prefers to build the robot “brain” with data, platforms, and integration solutions, while Samsung builds the “arm” with hardware, controls, and factory verification. The competition between the two sides has expanded to real-scenario performance, large-scale production capabilities, and commercialization.
