Summary of Key Points
In the field of embodied intelligence (robots that can perceive and act like humans), two companies, Yuzhu and Zhiyuan, are fiercely competing for a leading position. On one side is Yuzhu, which has just obtained A-share listing qualifications with an initial market value of approximately 42 billion yuan. On the other side is Zhiyuan, which is planning to list in Hong Kong with a target valuation of 40-50 billion Hong Kong dollars (about 35-43.3 billion yuan), clearly aiming to match Yuzhu's level. The two companies not only compete in the capital market but also claim different titles for "world's highest shipments." Zhiyuan claims to have the highest total shipments of humanoid robots, while Yuzhu insists it has the highest shipments of pure bipedal humanoid robots. More importantly, their business models are completely different: Zhiyuan relies on an ecosystem composed of local state-owned assets and investors to sell its robots, with revenue soaring from 300,000 yuan to 1 billion yuan in just three years; Yuzhu, on the other hand, sells standardized hardware to a diverse range of customers in the research, education, and developer communities. Who will emerge as the leader in this field depends on whether Yuzhu's hardware can be successfully integrated into real-world applications and whether Zhiyuan's model data can withstand the challenges of real environments.
I. Direct Comparison of Listing Valuations: The Contest for Dominance
As soon as Yuzhu obtained its A-share listing, Zhiyuan announced plans to list in Hong Kong with a target valuation of 40-50 billion Hong Kong dollars, nearly matching Yuzhu's market value of 42 billion yuan. This is no coincidence; Zhiyuan's investors have made it clear that they aim to compete with Yuzhu. Why do both companies place so much emphasis on valuation? In emerging industries, valuation represents not only the company's worth but also its influence in the industry. The one that lists at a higher valuation will attract more capital and talent and may even set the standards for the industry. This IPO showdown is like two heirs vying for succession, each seeking to become the legitimate representative of the embodied intelligence field.
II. Disputed Claims of "World's Highest Shipments": Hidden Secrets in the Statistics
At the beginning of this year, Omdia (a leading market research firm) claimed that Zhiyuan had the highest shipments of humanoid robots globally. Yuzhu immediately countered, stating that it had shipped over 5,500 pure bipedal humanoid robots, also ranking first. However, these claims are fundamentally different:
- Zhiyuan's statistics include a range of humanoid robots (bipedal, wheeled, quadrupedal, etc.), with 5,168 units shipped in 2025, accounting for 39% of the global total.
- Yuzhu only counts pure bipedal humanoid robots, with over 5,500 units sold.
Both companies are correct in their own ways, but they are both trying to prove to the market that their products are the best. In emerging industries, the title of "number one" can attract more customers and investors.
III. Divergent Business Models: Zhiyuan Relys on an Ecosystem, Yuzhu Sells Standardized Hardware
The two companies approach selling robots in very different ways, similar to how restaurants operate:
- Zhiyuan: Builds its business through an ecosystem. The secret to its rapid growth (3500-fold increase in revenue) lies in its partnership with local state-owned assets and investors. For example, in Zhuhai, Zhiyuan formed a joint venture that purchased robots for 12.73 million yuan; in Shaoxing, it collaborated with local governments and industry partners to secure orders worth 24.74 million yuan.
- Yuzhu: Sells standardized hardware. Its customers are spread across research institutions, educational organizations, and developers. Yuzhu's strategy is to develop reliable and affordable hardware (e.g., the G1 model starting at 99,000 yuan), which can be purchased directly from its website or JD.com. In the first nine months of 2025, its largest customer, JD.com, contributed only 3.54% of its revenue; the majority of orders came from universities, research institutions, and developers.
Yuzhu's approach focuses on making hardware (joints, motors, control algorithms) more mature and accessible as standardized products, similar to pre-assembled computers. The advantage is that this model is easy to replicate, but the challenge is that most robots are still in the research or demonstration phase and have not yet entered practical applications.
IV. Future Victory Points: Hardware maturity vs. Model Data Feedback Loop
Currently, both companies have their strengths, but the winner will depend on their ability to solve the core issue of making robots truly useful:
- Yuzhu's Strength: Solid hardware foundation. With ten years of experience in bipedal robots, Yuzhu can produce reliable and affordable robots at competitive prices (99,000 yuan for the G1 model), enabling them to perform tasks like running, jumping, and flipping. However, robots need more than just a physical body; they also require an intelligent "brain" to understand commands and complete tasks autonomously.
- Zhiyuan's Strength: Investment in data and models. Zhiyuan has developed the AgiBot dataset (a collection of 1 million training examples) and the GO-1 large model, which enables robots to learn from data. The more robots they deploy, the more data they collect, and the smarter the robots become, potentially expanding their use cases. However, this requires data collected in real-world scenarios, not just in laboratories.
V. Current Situation: Yuzhu Leads Temporarily, Zhiyuan as a Potential Disruptor
For now, Yuzhu appears to be the more established leader with mature hardware and a diverse customer base, reducing risks. Zhiyuan, on the other hand, is expanding rapidly through its ecosystem, but the sustainability of this model is uncertain, and its models need to be tested in real-world environments.
This competition for dominance is just beginning, and the IPOs are just the starting point. Who will win in the long run? It will depend on whether Yuzhu can transform robots from toys into practical tools that can be used in factories and homes, as well as whether Zhiyuan's models can overcome real-world limitations. The ultimate goal of embodied intelligence is for robots to function like humans, not just perform in laboratories.
(The entire analysis is written in plain language to make it understandable to non-financial readers.)