Summary of Key Points
In June 2026, the valuation of eight domestic humanoid robot companies exceeded 20 billion yuan, forming what could be called the "20 Billion Yuan Club." These eight companies can be categorized into three groups:
1. Companies that focus on selling complete hardware solutions (Yuzhu and Zhiyuan), which generate revenue through actual product shipments.
2. Companies that bet on the long-term value of the "brain" technology (such as Yinhe General), which invest in advanced models and data barriers to create a competitive advantage.
3. Companies that specialize in manufacturing key components (Lingxin Qiaoshou), which focus on developing robotic hands.
The article analyzes these companies from four perspectives: valuation logic, data collection methods, potential application scenarios, and the backgrounds of their founders and shareholders. It emphasizes that the 20-billion yuan valuation reflects capital's expectations for the future success of these companies, but their true value will ultimately be determined by practical indicators such as order volume, delivery performance, and the generalization ability of their models.
1. Valuation Logic: The Same 20 Billion Yuan, But Different Investments
The valuation logic of these eight companies reflects different assumptions about the future of humanoid robots:
- Investment in current sales potential (Yuzhu and Zhiyuan): These two companies are the only ones that can back their valuations with actual product shipments. Yuzhu achieved revenue of 1.699 billion yuan in 2025 and made a profit, while Zhiyuan shipped over ten thousand units, both backed by solid customer orders. However, if sales growth slows down, they may be perceived as traditional hardware companies (for example, FANUC has a market penetration rate of around 10 times, whereas Yuzhu’s is 24.7 times; the higher rate reflects the premium associated with humanoid robots). If sales growth fails to meet expectations, their valuations could decline.
- Investment in the potential of robot "brains" to become standard operating systems (Yinhe General and others): These companies have lower shipment volumes (in the hundreds) but have raised substantial funds, including investments from major players like Meituan, Alibaba, ByteDance, and Xiaomi. They believe that the technology behind these robot brains (model capabilities and data barriers) could become industry standards, similar to how Android and iOS became standard operating systems. The challenge is that progress in model development is less tangible than sales growth, and if investors lose patience, valuations could plummet.
- Investment in the strategic importance of key components (Lingxin Qiaoshou): This company focuses on manufacturing robotic hands and believes that as the humanoid robot market expands, all manufacturers will need such components. Their short-term revenue is more stable, but they face the risk of being replaced by companies that develop their own components (for example, Tesla's own Optimus hand).
2. Data Collection Methods: Simulation vs. Real-World Data
The development of robot brains relies on data training, and the eight companies use two main approaches:
- Simulation-based (Yinhe General): They create virtual environments in computers to train robots (with changing lighting conditions and object positions). This method is cost-effective and allows for rapid skill development, but the effectiveness of these skills in real-world scenarios remains uncertain.
- Real-world data-based (the other seven companies):
- Specialized data collection: Some companies use a "bodyless" approach, where humans operate the robots through wearable devices, reducing costs by 20% compared to traditional methods. However, this approach may introduce accuracy issues.
- Diverse data integration: Other companies combine various data sources, such as internet videos, wearable device data, and real-world robot data. Some also open-source their datasets to foster a broader ecosystem.
- Data accumulation through production: Companies like Yuzhu and Zhiyuan collect data while selling their products, which can help improve the generalization of their models across different scenarios.
3. Potential Application Scenarios: From Laboratories to Homes
The potential applications for humanoid robots determine the upper limit of their valuations. Currently, there are three main areas:
- Research and education (most mature but with limited growth potential): Yuzhu receives 73.6% of its revenue from university laboratories, and StarSeaMap serves over 150 research partners. This segment is crucial for company survival but may not be sufficient to support a valuation of 20 billion yuan due to limited market demand.
- Industrial production lines (highest potential but still in the early stages): Companies like QianXun have deployed robots in battery manufacturing lines, increasing efficiency significantly; Yinhe General has partnered with companies like Hikvision for semiconductor production line applications. However, these are still pilot projects.
- Home and commercial use (most challenging but with the greatest potential): Companies like Zhiyuan and 58到家 have collaborated on cleaning services, where robots perform tasks similar to human workers; Yinhe General is working on unmanned retail solutions; QianXun has partnered with JD.com for coffee delivery. These are experimental efforts and not yet ready for widespread adoption.
4. Founders and Shareholders: Why These Eight Companies Reached the 20-Billion Yuan Mark?
In the early stages of commercialization, the backgrounds of founders and shareholder resources play a crucial role:
- Founder expertise: Some founders come from academic backgrounds (Yinhe General, StarSeaMap) or have worked as scientists in large companies (Zhiyuan), bringing expertise in both technology and mass production.
- Growth momentum driven by investors: Companies with strong investor support, such as those backed by major tech firms like Meituan, Alibaba, and Tencent, can access valuable resources and market opportunities.
- Government support: Some companies (Yinhe General) have received investment from national AI funds, providing long-term funding and strategic guidance.
- Diverse investor portfolios: Investors include both government-backed entities and private investors with different priorities (market-oriented VCs for companies like QianXun and StarSeaMap), offering a balanced mix of resources.
Conclusion
The 20-billion yuan valuation is not an indication of immediate profitability but rather a "ticket" to enter the rapidly growing humanoid robot market. While Yuzhu and Zhiyuan have clear business models, their potential applications are still under question. Companies focusing on robot brains have complete technology but need time to prove their value. Lingxin Qiaoshou has a stable short-term outlook but faces the risk of being replaced by more advanced competitors. In the end, only companies that can meet key performance indicators (orders, delivery efficiency, profit margins, and model versatility) will truly justify such valuations. Time will reveal which ones succeed.