
Dobot is a leading enterprise in the collaborative robot sector, setting an A-share IPO record in 2026 by passing the listing review in just 86 days. The company has filed over 1,500 intellectual property applications, with more than 90% of key components self-developed, and has been recognized as a national-level "Little Giant" enterprise specializing in new technologies. In the first half of 2026, its revenue reached RMB 316 million, up 106.6% year-on-year.
In May 2026, Dobot unveiled its self-developed world action model, "Kongyi Dobot WAM", achieving an average completion rate of 99.25% across the four standard task suites of the LIBERO benchmark for embodied intelligence. Centered on the Kongyi large model, the company launched its "one brain, multiple bodies" embodied intelligence platform, enabling collaborative task execution based on a unified intelligent system.
On August 5, 2026, Dobot officially released its first embodied all-terrain humanoid robot, "LUMO" (DOBOT LUMO), standing nearly 1.3 meters tall. Powered by the Kongyi embodied large model, it achieves three core breakthroughs: multimodal emotional perception, 3D spatial understanding with proactive action, and autonomous learning with long-term evolution. Positioned for the home consumer market, "LUMO" is tailored to meet the diverse companionship needs of urban professionals, young families, and school-age children. Previously, Dobot had launched Dobot Atom (1.53 meters tall), a humanoid robot designed for industrial scenarios.
No official updates yet — the chain engine will fill this timeline from verified official sources.
Jensen Huang's daughter, Minshan Huang, visited Dobot's booth at the 2026 World Robot Conference to observe joint innovations in physical AI. She leads Nvidia's Omniverse and robotics product marketing efforts.
DPVR has partnered with a robot experience store in Qingdao to deploy its RoboPilot teleoperation and data collection solution. Using the E4 VR headset, users can control real robotic arms and dexterous hands in real time, while the system gathers data for embodied AI training.