Chinese humanoid robotics company Ubtech reported 2026 first half revenue of RMB 1.269 billion, up 104.2 percent year on year, with gross profit rising 160.9 percent to RMB 567 million. The company highlighted surging sales of full size embodied humanoid robots and progress on its self developed Thinker1.0 and Thinker WM AI models, while narrowing net losses to RMB 311 million.
This article aggregates reporting from 1 news source. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
Ubtech’s numbers are an early sign that humanoid robotics and embodied AI are starting to become real businesses, not just glossy demos. Revenue more than doubled, and sales of full-size humanoid robots grew over 14 times year on year, now nearly half the company’s top line. At the same time, Ubtech is talking up its Thinker1.0 foundation model and a world model variant, Thinker WM, which it claims rank at the top of multiple embodied benchmarks.
This matters for the race to AGI because embodied deployments force models to grapple with messy, physical environments instead of curated text and images. If Ubtech can ship hundreds or thousands of robots into industrial, commercial and even household settings, its world models will face a much richer distribution of edge cases than any synthetic dataset can provide.
Competitive pressure will follow. Tesla, Figure, Agility and a wave of Chinese startups are all chasing roughly the same prize: humanoid platforms that can perform useful work at scale. If Ubtech’s mix of in-house models and hardware momentum holds, it strengthens China’s position in embodied AI relative to US and European rivals, and increases the odds that some of the most interesting steps toward AGI-like generality will come from robots, not chatbots.



