The "billion yuan orders" in the embodied intelligence industry continue to be heard, and some doubts are gradually emerging. Some practitioners expressed their confusion: many humanoid robots are still struggling to perform in marathon races, how can they quickly enter practical scenarios and create a commercial closed loop? Are orders worth billions of yuan a genuine market demand or a hard to fulfill 'paper prosperity'? According to multiple interviews by reporters, some of the huge orders belong to framework agreements or intention orders, and coupled with the lack of transparency in the backgrounds and qualifications of both buyers and sellers, there is indeed a possibility of mixed content in their transactions. However, at the same time, we cannot ignore the catalytic effect of real demand on orders, including data collection, performance guidance, and simple handling and sorting in factories. In the process of maturing, any emerging technology will go through a stage from concept validation to large-scale commercial application. ”UBC CBO Tan Min told reporters that in 2024, UBC conducted intensive research on the needs of more than 10 intelligent manufacturing enterprises in the automotive and 3C industries, and found that they all have a demand for introducing humanoid robots in application processes such as handling, sorting, deployment, assembly, and testing. But the company will also clarify the feasibility of specific tasks based on the objective development progress of industry technology. Through these collisions, we can determine the range within which humanoid robots can truly function and achieve landing. On December 10th, Ubiquitous announced that it has recently signed a sales contract for humanoid robots with a domestic AI big model company, with a total amount exceeding 50 million yuan. The products are mainly the world's first industrial humanoid robot Walker S2 with independent battery swapping, which will be delivered within this year. As a listed company in the field of humanoid robots, Ubiquitous has disclosed the highest order amount in the industry. Excluding the full-size scientific research and education humanoid robot "Tiangongxing" and the small humanoid robot "AI Wukong", the total order amount for the Uber Walker series humanoid robots in 2025 has reached 1.3 billion yuan. Tan Min said that the above orders all come from clear industry demands, especially based on the urgent practical needs arising from the current transformation and upgrading of the manufacturing industry. In mid November, Walker S2 has started mass production and delivery, with the first batch of hundreds of units being put into frontline industrial applications, covering key areas such as automotive manufacturing, intelligent manufacturing, smart logistics, and embodied intelligent data centers. The reporter has seen from previous research that the embodied intelligent robot has been competent for some brief jobs in the factory. At the factory of Fulin Precision in Sichuan, the wheeled intelligent A2-W extends its narrow arms and drives multiple joints of the arms to flip, steadily lifting the material box on the ground. Wang Chuang, Partner of Zhiyuan Robotics and President of General Business Department, introduced to Shanghai Securities News reporters that the team had tried many scenarios before, including sorting, powder spraying, etc., but finally found that the loading and unloading boxes are currently a relatively suitable landing scenario. In April of this year, the success rate of transportation testing for Zhiyuan A2-W in the laboratory has reached 99.9%, compared to around 60% in the same period last year. In addition to going to the front line of the factory, performance guidance and data collection are also important "battlefields" for robots at present. The performance market goes without saying that at the beginning of the year, Yushu Technology's robots kicked out the market's enthusiasm for the embodied intelligence industry and made robots the top "traffic diversion" for major shopping malls and exhibitions. Starting from performance guided scenes with high attention and low technological barriers is a common law and practical choice for promoting new technologies, as well as a necessary starting point for expanding to industrial manufacturing and home service scenes in the future. Data collection is relatively unfamiliar to the public, but it itself is the cornerstone for robots to become more "intelligent". By continuously and systematically perceiving and recording the real world, robots can obtain massive and multidimensional environmental and task data. These data are the fuel that drives its' brain 'to learn and optimize, which can be used for algorithm training, scene understanding, system iteration, and more for robots. Guo Yandong, founder and CEO of Zhifang, analyzed to Shanghai Securities News reporters that the current order boom is essentially a positive response from the market to the use of embodied intelligent technology to solve practical business problems. Big models need real data feedback to evolve, and hardware must be polished in high-intensity applications. Only by making robots "work really" can "true intelligence" be cultivated. Despite the rapid growth of orders, the industry still faces hard constraints on mass production and delivery, as production capacity shifts from manual labor to mass production. According to Tian Feng, the director of the Fast Thinking and Slow Thinking Research Institute, multiple companies have disclosed framework agreements or intention orders with huge amounts but long delivery cycles. One important reason is that it will take another 1 to 2 years for factory production capacity to ramp up. At present, humanoid robots on the market are still mainly designed, assembled, debugged, and iterated manually by researchers or engineers in the laboratory. There are still a few enterprises capable of building mass production factories and producing humanoid robots on a large scale. Manual operation is difficult to ensure the ultimate precision required for precision parts, let alone meet the demands of large-scale production. Currently, multiple companies such as Songyan Power and Zhongqing Robotics are planning the construction of manufacturing factories to meet their expansion needs. Tan Min revealed that the current production capacity of industrial humanoid robots from Ubiquitous has reached 300 units per month, and it is expected that the annual delivery volume will exceed 500 units. Faced with the growing market demand, Ubiquitous has formulated a clear capacity ramp up plan: it is expected that by 2026, the annual production capacity of industrial humanoid robots will reach 5000 units, and further expand to 10000 units by 2027. We have great confidence in the richness and maturity of China's humanoid robot industry chain. From our understanding of the supply chain situation, most of Tesla and other companies' component partners are in China, concentrated in the Yangtze River Delta region and the Guangdong Hong Kong Macao Greater Bay Area. After years of accumulation, China has a large number of engineering talents in the fields of AI, robotics, and automation. These solid foundations will provide assistance in improving the delivery capabilities of Chinese humanoid robots. ”Tan Min said. From this, it can be seen that the "order tide" in the embodied intelligence industry is rooted in the real and urgent needs of manufacturing transformation and upgrading, and will ultimately be answered by time, market, and the stable performance of each robot on the production line. For the players involved, the significance of orders has long surpassed that of a single contract. Some investors from venture capital institutions believe that, on the one hand, orders are important chips for them to prove their commercial value to the capital market and obtain continuous "blood transfusion"; On the other hand, top enterprises can quickly accumulate real-world scenario data and build benchmark case barriers by seizing top customers, which is crucial for defining future technology routes and industry standards. (New Society)