Commercialization Year One? A Focus on China’s Humanoid Robot Industry in 2026

On December 8, Zhiyuan Robotics delivered its 5,000th mass-produced humanoid robot off the line; on December 13, Unitree announced the launch of a humanoid robot app store. Looking back at 2025—from U-Number Technology’s H1 dancing a yangge on the Spring Festival Gala stage at the beginning of the year, to ZQ’s T800 stepping into a boxing gym at year-end—alongside continued validation of the commercial value of related technology products, China’s humanoid robot industry may enter a “year one of commercialization” in 2026. How should we understand the posture of the humanoid robot industry? How should we perceive the development status and potential risks across the software–hardware industrial chain?

Scale of China’s Humanoid Robot Industry
In 2025, orders for China’s humanoid robot companies kept flowing in! According to estimates by the Humanoid Robot Scenario Application Alliance, China’s humanoid robot shipments in 2025 are expected to reach 20,000 units, a year-on-year increase of over 614%; market size will break through RMB 9 billion (approximately USD 1.25 billion), a year-on-year increase of 350%.

Morgan Stanley’s recently released The Morgan Stanley Robot Almanac sketches out an even grander long-term narrative. Although its estimate of China’s humanoid robot shipments is not as optimistic as some bullish forecasts, its top-level narrative centers on physical AI (i.e., robots + AI) as the core driving force of the Third Industrial Revolution. Robots will be widely applied in autonomous vehicles, household robots, humanoid robots, industrial robots, professional service robots, drones, and aerial vehicles (VTOLs), with the overall market set to enter a phase of explosive growth.

Focusing on the humanoid-robot segment, based on Morgan Stanley’s analysis and estimates, China demonstrates obvious leading advantages by virtue of early policy-side support, mature manufacturing capabilities on the industrial-chain side, and upstream–downstream raw materials and application scenarios.

Thinking on China’s Humanoid Robot Industrial Chain
For China’s humanoid robot industry, one ongoing topic of discussion is whether hardware (“movement”) should lead or software (“brain”) should lead. There is as yet no consensus on this point, and there is no right or wrong; both sides have achieved good progress in terms of commercialization and in capital markets.

For the hardware camp, results from an earlier AlphaWise Survey conducted by Morgan Stanley may well substantiate the pursuit of near-term landing. The survey targeted 86 industry executives—50% from manufacturing, 25% from services, and 25% from other industrial sectors. Fifty percent of respondents believe humanoid robots need only possess limited multi-tasking capabilities, while only 32% expect very general-purpose capabilities. On the other hand, 92% of respondents think the price of a humanoid robot should be RMB 200,000 and below (approximately USD 27,800 and below), of which 25% think it should be below RMB 50,000 (approximately USD 6,900), 28% chose RMB 50,000–100,000 (approximately USD 6,900–13,900), and 40% chose RMB 100,000–200,000 (approximately USD 13,900–27,800). In other words, lower prices and limited application scenarios (e.g., warehouse storage, production steps, customer service and retail) are an important foundation for being “willing to try using humanoid robots in the next three years.”

Matching this, upstream components generally benefit from demand multipliers exceeding a hundredfold brought about by the rapid development of the robotics industry. This is both “the most certain investment link” and contains enormous potential for scale-driven cost reduction. The Robot Almanac provides model estimates for major core components:

On the other hand, what the software camp pursues for ultimate usability of humanoid robots is the achievement of “four kinds of reachability.” First, mobility reachability: making it go wherever you want, including obstacle avoidance; second, manipulation reachability: being able to grasp whatever it can grasp or place an object at a designated location; third, semantic reachability: for example, taking an apple is for juicing; fourth, value and intelligence reachability. As Zhiyuan Robotics partner, head of the Embodied Business Unit, and executive dean of the research institute Yao Maoqing has expressed, AI capability will be the core link that differentiates the competitiveness of robot products; in the end, what robots compete on is AI capability, and a robot company that does not work on large models has no future.

Large models determine the generalization ability of humanoid robots—this is the fundamental reason humanoid robots are “humanoid,” and is also one of the present core moats of various robot manufacturers. A route that relies solely on empowerment from big-tech multimodal large models to run a layered end-to-end approach—if lacking self-research capability, and merely hoping that breakthrough AI advances in 3–5 years can be “seamlessly connected” to mass-produced robots—may easily be knocked down again by the waves of the times.

Admittedly, neither approach is right or wrong, but both carry risks. According to BCC Global Research, behind the current “high-spirited advance” of humanoid robot commercialization there may be a trend of relying on strategic cooperation and intention orders; that is, large purchase intentions come with specific technical conditions, and if early-stage technical verification fails to meet expectations, the large intention orders will automatically lapse and will not be disclosed. Thus, in the short term, both parties may benefit in terms of publicity volume or market-cap management, but the long-term business sustainability remains to be observed. On the other hand, R&D tackling of embodied-intelligence large models is by no means achieved overnight; given the scale or headcount limits of algorithm teams at humanoid-robot startups, whether they can replicate the engineering path of large language models (i.e., early breakthroughs by top players followed by rapid catch-up) remains unknown.

[Disclaimer]: The above content reflects analysis of publicly available information, expert insights, and BCC research. It does not constitute investment advice. BCC is not responsible for any losses resulting from reliance on the views expressed herein. Investors should exercise caution.