In recent years, the robotics industry has gradually become a hot concept sector in the Chinese capital market. Currently, there are 947 concept stocks in the A-shares market, with 252 of them related to humanoid robotics. However, in 2024, the market size of humanoid robotics in China was only 2.76 billion yuan. Industry insiders pointed out that the humanoid robotics industry is facing a dual dilemma of technology and commercialization. There is no robot company in A-shares that focuses on humanoid robotics, leading the industry into a false prosperity with capital arbitrage quietly surging.
According to data from Tonghuashun, as of Wednesday (July 2), there are a total of 947 listed companies related to the robotics concept in the A-shares market, equivalent to one out of every six A-share companies being involved in the robotics business, with 252 of them focusing on humanoid robotics.
The fate of these nearly one thousand robotics companies could be very pessimistic.
On June 28, at the 2025 Huaying Capital Annual Conference, Wang Qian, the founder and CEO of Yinshu Robotics, stated that based on the development history of the mobile phone and automobile industries, there are now only six or seven major manufacturers in the mobile phone market, while the automobile market still has dozens of them. Robotics should fall somewhere in between. The global market is expected to eventually shrink to around twenty major manufacturers.
The ultimate outcome could be similar to the new energy vehicle industry, where the majority of companies fail. From 2014 to 2019, there were once over 60 new players in the Chinese automotive market, but currently, only six remain operating steadily – including companies like Li Auto, XPeng, Leapmotor, Xiaomi, NIO, and Jidu.
A recent article by the 20-year-old automotive e-commerce platform, “Gaishi Cars,” pointed out that the humanoid robotics industry in 2024 is experiencing a “false prosperity.” Companies are announcing mass production plans to cater to capital trends. While Zhiyuan Robotics claimed to be “mass producing a thousand units in 2025,” its products have not yet passed the reliability tests of partners; Agility’s Digit robot, though expanding production, faces delivery delays due to shortages of core components. Zhang Yu, the founding partner of Qingzhi Capital, warned that “mass production is just digital games; a lack of scenario-verified production capacity will become a toxic inventory.”
This kind of “false prosperity” is still ongoing in 2025. Companies like Qianxun Intelligence, Songyan Power, and Zhiyuan Robotics have recently secured new rounds of financing, and the financing wave in the humanoid robotics field is booming. However, on March 28, Zhu Xiaohu, a partner at Gobi Partners, issued a warning about the humanoid robotics industry, questioning, “Who would spend hundreds of thousands to buy robots to work?” and publicly announced a “bulk exit from the humanoid robotics race.”
Zhu Xiaohu’s comments “exposed the emperor’s new clothes”: in 2024, the market size of humanoid robotics in China was only 2.76 billion yuan, yet the amount of financing reached 5 billion yuan, with valuations far exceeding actual demand. However, there are differing views. Zhang Ying from Vertex Ventures believes that in the robotics field, having some froth is quite normal during the “big race” where many flowers bloom.
As the industry froth emerges, capital arbitrage is also quietly happening in the background. Many loosely related companies are capitalizing on the humanoid robotics trend through various means, pushing up stock prices through shady practices.
Yu Fenghui, a senior researcher at Panghu Think Tank, stated that some companies announce high investments in robotics research projects without actually allocating corresponding resources, while others release vague cooperation statements claiming strategic partnerships with well-known robotics companies to boost market expectations and stock prices.
A senior investment banker at a brokerage firm in South China noted that most companies’ humanoid robotics businesses are still in the initial stages of development, generating little revenue growth, and the development stage of the business does not align with the actual valuation; even if there are collaborations or investment relationships with popular robotics companies, the actual business dealings between the two parties might be limited, and these robotics companies have not yet gone public. “Some companies’ stock price increases are clearly unreasonable, driven more by irrational speculation.”
Renowned business consultant and corporate strategy expert Huo Hongyi revealed common practices within the industry. He said, “In situations where there are so-called ‘strategic partnerships’ with leading robotics companies, funds are frequently exchanged in specific stocks before signing agreements. After the agreement is signed and the news is released, the stock price is quickly pushed up, followed by rapid shipping.”
Both in terms of technology and commercialization, humanoid robotics are facing bottlenecks. An article by “Gaishi Cars” mentioned that the technical challenges of humanoid robotics are far more significant than commonly imagined. For example, bipedal walking involves complex issues like dynamics control, real-time balance algorithms, joint torque distribution, among others. The research team at Boston Dynamics revealed that achieving stable walking alone took ten years, while functionalities like jumping and backflips require thousands of calculations per second for repeated iterations.
A more significant bottleneck lies in cognitive capability. Jiang Lei, from the Chinese National Humanoid Robotics Innovation Center, indicated that current robots lack the ability for “common-sense reasoning.” Wang He, the founder of Galaxy General, admitted that “making a robot understand simple instructions like ‘fetch a drink from the second shelf of the fridge’ requires the fusion of visual recognition, spatial memory, and task planning, which currently lacks mature solutions.”
The commercialization path of humanoid robotics is also not yet clear. In domestic settings, they face ethical and safety controversies; in the educational field, the practicality of humanoid robotics is also highly questioned.
The cost issue further exacerbates the dilemma. The BOM (Bill of Materials) cost for Optimus from Tesla is around $100,000, and domestic enterprises attempting to lower prices by reducing the number of sensors have led to degraded functionalities. Zhang Yu from Qingzhi Capital pointed out, “Robots priced at 30,000 yuan can only perform gymnastic routines but cannot actually accomplish real tasks.”
The Securities Times reported that the humanoid robotics industry is still in the research and development stage of “from zero to one,” and there is still a long way to go before actual industrial implementation. Therefore, there is currently no A-share company whose main business is the R&D and manufacturing of humanoid robots.
During the recent NVIDIA GPU Technology Conference hosted by NVIDIA, Pieter Abbeel, a professor at the Department of Electrical Engineering and Computer Science at the University of California, Berkeley, mentioned that there are currently no real humanoid robots in the world.
The report indicates that there are three common paths for robotics concept stocks to enter this realm. One is to act as a core component supplier, entering the humanoid robotics supply chain. Another is to partner with top-tier robotics companies in the primary market to engage in technical cooperation and ecosystem layout within the robotics business. The third path is taken by many traditional companies expanding into the robotics field.