Summary of Key Points
As the first humanoid robotics company listed on the A-share market, Yushu Technology saw its stock price soar by six times on the day of its listing. By leveraging China's supply chain and exceptional cost control, it has become a leader in the industry, reducing the price of humanoid robots by over 400,000 yuan within two years while increasing its gross profit margin from 44% to 60%. This has also prompted upstream component manufacturers to shift from custom production to standardized manufacturing. However, Yushu has a significant weakness: the “brain” of the robots (the autonomous algorithm models) is not yet mature. The robots currently sold are mainly purchased by research institutions for testing purposes and are far from being ready for use in actual factory operations.
I. Yushu’s Secret to Becoming a Leader in the Industry: Extreme Cost Efficiency, with Help from China’s Supply Chain
Yushu’s founder, Wang Xingxing, is known for his frugality—even when the company is financially stable, he still cares about who pays for hotel expenses during meetings. This focus on cost, combined with decades of experience in Chinese manufacturing, has enabled Yushu to become the only manufacturer in the industry that both sells a large volume of products and generates substantial profits.
- Innovation Driven by Need: Due to the company’s small size in its early stages, suppliers were reluctant to customize components, forcing Yushu to develop its own motors, joints, and lidar systems. This led to better cost control. For example, the price of humanoid robots dropped from 590,000 yuan in 2023 to 160,000 yuan in 2025, yet the gross profit margin increased because costs were reduced more quickly than prices.
- The Supply Chain Adapts: Upstream companies have changed their production methods to meet Yushu’s demands, shifting from custom parts to standardized products (such as those produced by Zhuoyu Technology, which now offers three types of standard components covering 95% of robot needs). The flexibility of China’s supply chain has allowed Yushu to mass-produce robots while maintaining profitability, thus establishing its dominant position in the industry.
II. Changes in the Upstream Supply Chain: From Customization to Industrial Mass Production, but With Challenges
Yushu’s orders are shaping the development of the upstream supply chain, although there are still unresolved issues:
- The Transition from Handmade to Industrialized Production: Robot components were previously custom-made; now they are becoming standardized, marking a shift from a small-scale manufacturing approach to an industrial one. Established companies like Audiwei, which has been in the sensor industry for 27 years, have entered the robotics market and sold tens of millions of sensors.
- Challenges with High-End Components: High-end components (such as high-precision six-axis force sensors and AI chips) still need to be imported. Additionally, robot joints often suffer from heat dissipation issues (early models required watering; now fans are used, but they can break easily), which poses challenges for applications in continuous operations.
III. The Missing “Brain”: Robots Can Move But Cannot Think—A Major Barrier Despite Investment
Yushu’s hardware is world-leading, but its “brain” (the AI models) is still underdeveloped:
- Unbalanced R&D Focus: In previous years, Yushu focused most of its R&D on the hardware (e.g., motion control). Only in 2024 did it begin investing in AI model development. The 4.2 billion yuan raised from this IPO will be half used for model research and development.
- Barriers to Progress: Collecting data for robot “thinking” requires real-world physical information (e.g., friction and fluid dynamics), which is extremely costly (a large company spent millions just on data collection). AI chips rely on NVIDIA, and domestic alternatives lack sufficient computing power and are expensive. Currently, the intelligence level of robots is similar to that of Level 3 autonomous vehicles—capable of simple tasks in controlled environments but struggling in complex ones.
IV. Who Are the Buyers? Mostly Research Institutions—Still Far from Practical Use
China accounts for 97% of global humanoid robot shipments, but the customer base indicates that the industry is still in its early stages:
- Mainly Research-Driven: In 2025, 73.6% of Yushu’s revenue came from research and education applications, with only 9% from industrial use. The top five customers together account for just 12%, indicating a lack of large-scale industrial clients.
- Similar to Early PCs: Today’s robots serve as development platforms—companies in the U.S. that develop AI algorithms use Yushu’s robots (since they cannot afford Tesla’s). Similar to how people bought PCs to develop software, practical applications in factories await breakthroughs in AI models and data collection.
Conclusion
Yushu has demonstrated that Chinese manufacturing can produce affordable humanoid robots for global markets. However, to make robots truly useful in industrial settings, the “brain” (AI models) must be developed. This is not only a challenge for Yushu but also for the entire industry, requiring advancements in AI models, data collection, and chip technology. Yushu’s current position is similar to that of early PC manufacturers: it has focused on building high-quality hardware and is waiting for the software ecosystem to mature.