LG CNS has made a bold move to bolster its humanoid robot hardware capabilities, ramping up efforts to commercialize physical artificial intelligence (AI).
The company announced on Tuesday that it has become the first South Korean firm to make a strategic investment in Dexmate, a U.S.-based robotics company. This investment was executed through LG Technology Ventures, LG’s corporate venture capital arm.
Dexmate, headquartered in Silicon Valley, has earned recognition for its high-quality, high-performance humanoid robots. These robots have been adopted as standard research hardware by leading global robotics brain developers. A key feature of Dexmate’s robots is their use of wheels instead of legs, maintaining human-like task capabilities while ensuring stable, long-duration operations.
The robots are equipped with a wheel-based lower body, dual arms optimized for high-speed tasks, and a head with vision sensors for environmental awareness. They boast over 36 degrees of freedom, enabling precise two-handed collaborative work. The robots can handle loads of up to 15 kg (about 33 lbs) with their arms and operate for more than 20 hours on a single charge. Their stable wheeled base makes them more suitable for various industrial settings, such as logistics centers and manufacturing plants, compared to traditional bipedal humanoid robots.
This investment allows LG CNS to round out its robot hardware lineup, now including bipedal humanoids, quadrupedal robots, and wheeled humanoids. Building on this diverse portfolio, the company plans to launch a full-stack reactive extensions (RX) service, combining three critical elements for humanoid robot commercialization: hardware, Robot Foundation Model (RFM), and an operational and learning platform.
LG CNS is actively developing its own platform for robot operation and learning while expanding partnerships and investments in the robotics sector. In June last year, the company invested in and partnered with Skilled AI, a U.S.-based robotics brain development firm, to enhance industry-specific RFMs.
The company is also conducting numerous proof-of-concept (PoC) projects with robots trained on data from logistics, distribution, and manufacturing environments. These robots are currently being tested in real-world scenarios, such as loading and sorting items in logistics centers and inspecting the assembly and quality of ships.
Lee Joon-Ho, Executive Vice President of LG CNS’s Smart Logistics & City Division, explained that this investment is a strategic move to seamlessly integrate robot hardware, RFM, and platforms. It will enable large-scale robot deployment and rapid adoption in industrial settings. The goal is to pioneer the physical AI era by demonstrating a humanoid robot business model that’s immediately applicable in real-world scenarios, moving beyond mere technological validation.
