Physical AI theme strengthens, boosting humanoid robot concept stocks; technological maturity and commercialization remain to be verified
The physical AI theme strengthened in the A-share market, driving the humanoid robot concept stock index up more than 5% during intraday trading on October 28, with multiple stocks such as Fantuo Digital Creative and Fulei New Materials hitting the daily limit up for consecutive sessions. Several companies subsequently issued announcements warning of overheated market sentiment and risks of irrational speculation. Unitree Technology stated that the humanoid robot industry is still in the early commercialization stage, and that physical AI still faces multiple challenges, including perception accuracy and motion control stability, in moving from laboratory demonstrations to large-scale deployment.
Body · AI translation
Recently, the physical AI theme has gained momentum in the A-share market, with multiple related concept stocks hitting the daily limit. Afterwards, some companies successively issued announcements on abnormal trading fluctuations, with one company warning of "overheated market sentiment and irrational speculation, the possible existence of irrational trading behavior, and relatively high trading risks."
After a brief pullback on October 27, the physical AI theme heated up again on October 28, with the related humanoid robot concept stock index rising more than 5% during intraday trading.
Shen Meng, executive director of Chanson Capital, analyzed to reporters that this market movement is a fluctuation driven by individual industry events rather than value fundamentals, with investor sentiment being the main driver. He believes that physical AI is still in the stage of academic exploration and R&D, and commercial scenarios lack sufficient attempts.
The "Embodied Intelligence Development Report (2026)" released by the China Academy of Information and Communications Technology mentions that physical AI aims to break through the limitations of traditional models that "only represent and lack physical constraints," referring to embedding physical laws into the entire process of perception, reasoning, generation and action, so that artificial intelligence possesses greater authenticity, reliability and deployability.
Over the past two years, physical AI has also become a focus of industry attention. In his GTC speech on October 21, similar to views he had expressed previously, Nvidia CEO Jensen Huang once again mentioned that "physical intelligence is the next wave of AI," and said that the physical AI application with the largest shipment volume is autonomous driving.
The day after Jensen Huang's speech (October 22), Unitree Robotics founder Wang Xingxing said at a conference that with the continuous iteration of large models, physical AI may see more progress, possibly in one to two years at the fastest, or two to three years at the slowest.
Shen Meng also added that AMD's previous high-priced acquisition of Fei-Fei Li's world model company likewise played a catalytic role. In addition, some local governments have also issued policies to support the development of physical AI and related industries.
Reporters noted that related concept stocks have had strong momentum for several days, among which data collection, simulation platforms and others have received more attention. From October 19 to October 28, Fantuo Digital Creative picked up three daily limit-ups and set a new share price high of 70.78 yuan per share on October 26; as of October 26, its cumulative gain over six trading days had doubled. Century Hengtong previously picked up one daily limit-up, and on October 26 its gain reached 17.74% again. Orbbec's share price rose, with a cumulative gain of nearly 50% over seven trading days.
It is worth noting that the humanoid robot sector has also received a certain boost, with the index rising more than 8% cumulatively since October 19. Under the sub-concepts, companies related to roller screws, reducer bearings, electronic skin and others are favored. Daye Co., Fulei New Materials and many other stocks rose for several consecutive days, with Fulei New Materials even notching five daily limit-ups. Unitree Robotics' share price also showed a corrective rebound, with its market value returning above 250 billion yuan. Shangwei New Materials also recorded a daily limit-up on October 26 and touched the daily limit intraday on October 28.
Prior to this, the humanoid robot sector's index had overall declined since late September. In terms of individual stocks, Unitree Robotics' share price came under pressure for a time after listing and hit a stage low of close to 400 yuan per share on October 15.
Funds advancing and retreating in the short-term market movement
Choice data shows that in this short-term market movement, some related concept stocks also appeared on the Dragon and Tiger List.
For example, Fantuo Digital Creative showed active trading for several days. Specifically, Dragon and Tiger List data shows that on October 20, the early stage of the market movement, the stock had total net buying of more than 75 million yuan. Among the top five buying seats, institutional dedicated seats occupied four. In contrast, among the top five selling seats, Guoxin Securities Beijing Branch and Lianchu Securities Hangzhou Huancheng North Road occupied two.
In the following trading days, divergences in funds emerged. From October 21 to 23, selling funds predominated, with total net buying of -10.82 million yuan, -22.54 million yuan and -15.06 million yuan respectively. Especially on October 23, the sell amounts of two institutional dedicated seats themselves were both higher than their buy amounts, as they chose to cash out on the daily limit-up day. Hot money seats began to become an important force on the buying side, with Dongguan Securities Nanjing Branch and CITIC Construction Investment Fuzhou Dongjie replacing institutions as the buyers that day. By the 26th, there were no institutional dedicated seats among the top five buying seats, and institutions still sold more than 100 million yuan that day; entering the 27th, buying and selling tended toward balance.
Overall, institutions were the initiators of the company's early market movement, and hot money subsequently participated.
It is worth noting that in June this year, physical AI concept stocks also experienced a wave of gains. At that time, Tianyu Digital Science picked up multiple daily limit-ups and then issued an announcement clarifying that it had no physical AI business. According to media reports, hot money and institutions also participated at that time.
As for this market movement, a number of listed companies subsequently issued announcements warning of related risks and "pouring cold water" on it.
On October 23, Fantuo Digital Creative said that the company's stock trading had shown severe abnormal fluctuations, with a relatively large short-term gain and a relatively high turnover rate, and that there was overheated market sentiment and irrational speculation. It is understood that the company's main business is 3D visualization products and services, integrated digital services, AI 3D digital twin products and solutions, and others. The latest financial report shows that in the first three quarters of this year, Fantuo Digital Creative remained in a loss-making state, with a net loss attributable to shareholders of 123 million yuan.
Daye Co. even issued multiple announcements clarifying its business situation. "The company's existing main business remains focused on the R&D, production and sales of bead wire, steel cord and rubber hose wire. The company's robot tendon rope products are currently still in the R&D and sample submission stage, have not yet achieved mass production, and have not generated substantive revenue," the company said.
Unitree Robotics also stated in its announcement that the humanoid robot industry is still in the early commercialization stage, and that there is uncertainty in technology maturity, cultivation of downstream market demand, and supporting capacity of the industrial chain. Physical AI technology still faces multiple challenges in moving from laboratory demonstrations to large-scale implementation in real scenarios, including perception accuracy, motion control stability, data accumulation and software-hardware coordination, and it still takes time to verify the maturity of related technical paths and business models.