#股票交易分享挑战 Will Unitree Technology, valued at ¥60 billion, be too expensive?
Unitree Technology’s IPO valuation is approximately ¥61 billion, while its valuation on the OTC market was once inflated to more than ¥200 billion. However, less than 3% of its current humanoid robot revenue genuinely comes from production scenarios such as manufacturing, inspection, and logistics. The price the market is currently assigning to Unitree is clearly not based on today’s performance, but on the story of robots fully entering factories. The question is: how far away is that day?
Over the past two years, humanoid robots have been one of the sexiest stories in China’s technology industry (possibly the sexiest). And Unitree Technology is the most dazzling company in this story.
Robots running, fighting, dancing, and even appearing on the CCTV Spring Festival Gala—from technical capabilities to brand recognition, Unitree Technology has become a representative of China’s humanoid robot industry. Now, it is finally going public, and the capital market has naturally responded with the greatest enthusiasm. The offering price of 150.80 yuan implies a market capitalization of approximately ¥61 billion and an offering P/E ratio of 219x. Overseas OTC markets and related pre-market contracts even at one point implied a valuation of more than ¥10k. What does a P/E of 219x mean? During the same period, the average static P/E ratio for the general equipment manufacturing industry was only around 38x.
The question is, how many robots would a robotics company have to sell to support a ¥200 billion valuation? Figure 1: CCTV’s report on Unitree.
I. ¥60 billion—what is the market buying?
Of course, directly comparing a robotics company with traditional equipment manufacturers based on P/E is not particularly meaningful. People are spending ¥60 billion to buy Unitree because they are betting on the future.
China has hundreds of millions of industrial workers. As AI foundation models, motion control, and robot hardware continue to mature, humanoid robots will eventually enter factories, warehouses, and commercial settings, performing large amounts of repetitive work such as handling, assembly, and inspection. At that point, a robot will no longer be equipment costing hundreds of thousands of yuan, but could become a new form of “labor.”
The business of selling robots will become the business of selling labor. The valuation space does indeed offer considerable room for imagination. But a closer look at Unitree Technology’s current revenue mix reveals something interesting. In the first nine months of 2025, 73.6% of Unitree’s humanoid robot revenue came from research and education, 17.39% came from commercial consumption and other scenarios, and genuine industrial applications accounted for only 9.01%. Breaking down that 9% further, approximately 70% of industrial applications consisted of corporate tours, mainly involving greeting visitors, exhibition hall explanations, route guidance, and interactive Q&A. Revenue genuinely coming from smart manufacturing, smart inspection, and logistics delivery was only approximately ¥15.7 million. In other words, revenue genuinely related to “robots entering factories to work” accounted for less than 3% of humanoid robot revenue.
That is somewhat awkward. Because in the current humanoid robot—or embodied intelligence—industry, the biggest narrative is neither dancing nor fighting, but precisely entering factories.
II. Shipment volume ≠ commercialization.
SemiAnalysis previously estimated that Unitree’s robot shipments had exceeded 5,500 units in 2025. That sounds like a lot, but only around 250 may have actually entered industrial application scenarios. Where did the rest of the robots go? To universities, laboratories, AI companies, technology companies, and exhibition halls. Of course, universities buy robots to research embodied intelligence, AI companies buy robots to train models, and technology companies may buy robots to place in exhibition halls. These are all genuine sources of demand. But they are entirely different business models from “a factory purchasing 10k robots to replace workers.”
The former is closer to high-end research equipment: it has high gross margins, but extremely limited volume. The latter could become a truly mass-market industrial product, with room for shipments in the millions or even tens of millions. The capital market is clearly valuing Unitree according to the second story, while the revenue structure currently disclosed by Unitree remains rooted in the first.
A high valuation is not necessarily unacceptable. When technology companies are just entering a phase of explosive growth, P/E does indeed have limited reference value. As long as revenue and profits continue growing rapidly, even an outrageous valuation today may later be digested by performance. The problem is that Unitree’s growth has also begun to show some changes worth noting. In the first quarter of this year, Unitree’s revenue was approximately ¥423 million, up 68.49% year on year. That figure remains high, but compared with growth of more than 300% during the same period previously, it has clearly slowed. Even more noteworthy is net profit excluding non-recurring items. During the same period, net profit excluding non-recurring items fell from approximately ¥84.84 million to ¥40.25 million, a year-on-year decline of 52.55%. There are reasonable explanations for this: Unitree continued to invest heavily in embodied intelligence foundation models, motion control algorithms, and robot body structures, with R&D expenses alone increasing by approximately ¥38 million year on year.
For a rapidly growing technology company, spending more on R&D is not necessarily a bad thing. But for investors, the problem remains significant: a company with rapidly growing profits can gradually digest a high valuation through its performance; what can a company whose profits have fallen 50% rely on to maintain a 200x P/E ratio over the long term?
III. Unitree’s dilemma
Unitree Technology’s biggest problem now is not insufficiently advanced technology. Being able to develop quadruped and humanoid robots to their current level has already demonstrated Unitree’s capabilities. What truly remains to be verified is Unitree’s ability to create economic value on a large scale through its products. If humanoid robots really enter factories rapidly over the next few years, with demand growing from several thousand units to hundreds of thousands or even more, while Unitree can maintain its current technological edge, then looking back several years from now, today’s ¥60 billion—or even ¥200 billion—valuation may not seem outrageous. But one thing is clear: Unitree’s robots can already run, jump, and throw punches. Next, they do not need to perform another backflip; they need to quickly find a factory and go to work. UNITREE$UNITREE
Unitree Technology’s IPO valuation is approximately ¥61 billion, while its valuation on the OTC market was once inflated to more than ¥200 billion. However, less than 3% of its current humanoid robot revenue genuinely comes from production scenarios such as manufacturing, inspection, and logistics. The price the market is currently assigning to Unitree is clearly not based on today’s performance, but on the story of robots fully entering factories. The question is: how far away is that day?
Over the past two years, humanoid robots have been one of the sexiest stories in China’s technology industry (possibly the sexiest). And Unitree Technology is the most dazzling company in this story.
Robots running, fighting, dancing, and even appearing on the CCTV Spring Festival Gala—from technical capabilities to brand recognition, Unitree Technology has become a representative of China’s humanoid robot industry. Now, it is finally going public, and the capital market has naturally responded with the greatest enthusiasm. The offering price of 150.80 yuan implies a market capitalization of approximately ¥61 billion and an offering P/E ratio of 219x. Overseas OTC markets and related pre-market contracts even at one point implied a valuation of more than ¥10k. What does a P/E of 219x mean? During the same period, the average static P/E ratio for the general equipment manufacturing industry was only around 38x.
The question is, how many robots would a robotics company have to sell to support a ¥200 billion valuation? Figure 1: CCTV’s report on Unitree.
I. ¥60 billion—what is the market buying?
Of course, directly comparing a robotics company with traditional equipment manufacturers based on P/E is not particularly meaningful. People are spending ¥60 billion to buy Unitree because they are betting on the future.
China has hundreds of millions of industrial workers. As AI foundation models, motion control, and robot hardware continue to mature, humanoid robots will eventually enter factories, warehouses, and commercial settings, performing large amounts of repetitive work such as handling, assembly, and inspection. At that point, a robot will no longer be equipment costing hundreds of thousands of yuan, but could become a new form of “labor.”
The business of selling robots will become the business of selling labor. The valuation space does indeed offer considerable room for imagination. But a closer look at Unitree Technology’s current revenue mix reveals something interesting. In the first nine months of 2025, 73.6% of Unitree’s humanoid robot revenue came from research and education, 17.39% came from commercial consumption and other scenarios, and genuine industrial applications accounted for only 9.01%. Breaking down that 9% further, approximately 70% of industrial applications consisted of corporate tours, mainly involving greeting visitors, exhibition hall explanations, route guidance, and interactive Q&A. Revenue genuinely coming from smart manufacturing, smart inspection, and logistics delivery was only approximately ¥15.7 million. In other words, revenue genuinely related to “robots entering factories to work” accounted for less than 3% of humanoid robot revenue.
That is somewhat awkward. Because in the current humanoid robot—or embodied intelligence—industry, the biggest narrative is neither dancing nor fighting, but precisely entering factories.
II. Shipment volume ≠ commercialization.
SemiAnalysis previously estimated that Unitree’s robot shipments had exceeded 5,500 units in 2025. That sounds like a lot, but only around 250 may have actually entered industrial application scenarios. Where did the rest of the robots go? To universities, laboratories, AI companies, technology companies, and exhibition halls. Of course, universities buy robots to research embodied intelligence, AI companies buy robots to train models, and technology companies may buy robots to place in exhibition halls. These are all genuine sources of demand. But they are entirely different business models from “a factory purchasing 10k robots to replace workers.”
The former is closer to high-end research equipment: it has high gross margins, but extremely limited volume. The latter could become a truly mass-market industrial product, with room for shipments in the millions or even tens of millions. The capital market is clearly valuing Unitree according to the second story, while the revenue structure currently disclosed by Unitree remains rooted in the first.
A high valuation is not necessarily unacceptable. When technology companies are just entering a phase of explosive growth, P/E does indeed have limited reference value. As long as revenue and profits continue growing rapidly, even an outrageous valuation today may later be digested by performance. The problem is that Unitree’s growth has also begun to show some changes worth noting. In the first quarter of this year, Unitree’s revenue was approximately ¥423 million, up 68.49% year on year. That figure remains high, but compared with growth of more than 300% during the same period previously, it has clearly slowed. Even more noteworthy is net profit excluding non-recurring items. During the same period, net profit excluding non-recurring items fell from approximately ¥84.84 million to ¥40.25 million, a year-on-year decline of 52.55%. There are reasonable explanations for this: Unitree continued to invest heavily in embodied intelligence foundation models, motion control algorithms, and robot body structures, with R&D expenses alone increasing by approximately ¥38 million year on year.
For a rapidly growing technology company, spending more on R&D is not necessarily a bad thing. But for investors, the problem remains significant: a company with rapidly growing profits can gradually digest a high valuation through its performance; what can a company whose profits have fallen 50% rely on to maintain a 200x P/E ratio over the long term?
III. Unitree’s dilemma
Unitree Technology’s biggest problem now is not insufficiently advanced technology. Being able to develop quadruped and humanoid robots to their current level has already demonstrated Unitree’s capabilities. What truly remains to be verified is Unitree’s ability to create economic value on a large scale through its products. If humanoid robots really enter factories rapidly over the next few years, with demand growing from several thousand units to hundreds of thousands or even more, while Unitree can maintain its current technological edge, then looking back several years from now, today’s ¥60 billion—or even ¥200 billion—valuation may not seem outrageous. But one thing is clear: Unitree’s robots can already run, jump, and throw punches. Next, they do not need to perform another backflip; they need to quickly find a factory and go to work. UNITREE$UNITREE
