Dobot Unveils LUMO: ~1.3 m Companion Humanoid for Home, Campus, and Grass
Collaborative-arm maker DOBOT has stepped further into consumer-facing bipeds. On August 6, 2026, 36Kr Europe reported the launch of DOBOT LUMO, described as a full-scenario “embodied all-terrain” humanoid aimed at homes, campuses, and outdoor recreation rather than factory totes.
What the launch claims
According to the 36Kr write-up, LUMO is nearly 1.3 meters tall and is shown walking on lawn, sand, and cobblestone, switching between outdoor and indoor scenes. The piece frames the product as a multi-role companion: sports playmate outdoors, emotional companion at home, education partner for school-age kids, and stage performer for commercial events.
Brand marketing director Xie Kaixuan is quoted defining “embodied all-terrain” across four axes: multi-space mobility (home, mall, campus), multi-role identity, multimodal vision/voice/motion capability, and long-term learning with the user. The stack is said to run on DOBOT’s self-developed Kongyi Embodied Large Model, with claims of multimodal emotion perception, 3D spatial understanding with navigation and obstacle avoidance, and closed-loop learning from real-scene data.
Context from the same coverage
The article positions LUMO for urban office workers, young families, and K–12 / research users who want an embodied teaching platform with secondary-development hooks (motion control, SLAM, vision, HRI). It does not publish a public U.S. MSRP or shipping calendar in the English 36Kr piece — those numbers stay unstated in our sources, so we leave them out.
DOBOT already sells industrial-leaning humanoids such as Atom (listed elsewhere with China/US store pricing). LUMO is a different product pitch: shorter stature, companion aesthetics, and outdoor gait demos rather than assembly-line dexterity.
A Human’s Take
I care less about “world’s first” marketing and more about what the video actually shows: continuous walking on real outdoor surfaces and a clean cut between sport and home modes. Emotion-recognition claims and backflip clips are easy to over-index. What I want next is runtime numbers, fall recovery, and whether LUMO can do a multi-hour campus shift without a minder. Until DOBOT publishes those, treat LUMO as a capable demo platform with a clear consumer target — not a proven home deployment.