Unitree's valuation may be overinflated due to high subscription demand and market hype.

Original | Odaily Planet Daily (@OdailyChina)

Author | jk

On the evening of August 6, Unitree Technologies announced its IPO price on the STAR Market at RMB 150.80 per share, corresponding to an offering market capitalization of RMB 60.993 billion. On the day of online subscription, August 10, 9.78 million retail investors participated, resulting in an initial effective subscription multiple of 8,288.82 times—the lowest oversubscription rate in STAR Market history. On Hyperliquid’s perpetual futures platform, Unitree’s pre-market quote surged to $90, translating to a market capitalization exceeding RMB 240 billion—four times its IPO valuation of RMB 61 billion. Together, these figures paint a picture of near-universal frenzy surrounding the IPO. Yet behind the excitement, an increasing number of voices are questioning whether this price has already priced in too much speculation.

Today, Odaily Planet Daily will take you through five reasons to short the rally.

On August 6, Yushu Technology announced its offering price at RMB 150.80 per share, corresponding to an offering market capitalization of approximately RMB 60.99 billion. What truly startled the market was its price-to-earnings ratio of 219.23; for comparison, the average static P/E ratio for the general equipment manufacturing industry over the past month, as published by China Securities Index, was only around 38.56, meaning Yushu’s valuation is roughly 5.7 times the industry average.

CCTV's report on Unitree. Source: CCTV

Some investors have jokingly referred to this P/E ratio as the "dream ratio," meaning that the price paid is not for current profitability, but for a distant dream.

On overseas secondary market platforms such as Hiive, Unitree's share price reached as high as $61.63 between August 10 and 11, implying a total valuation of nearly $22.4 billion—more than double the valuation at the offering price.

On Hyperliquid, the situation is even more extreme. As of the time of writing, the pre-market perpetual contract UNITREE-USDC for Unitree Robotics was quoted at around $90, implying a market capitalization of approximately $35.47 billion based on 404 million total shares outstanding—equivalent to about RMB 239 billion, nearly four times its issued market capitalization of RMB 60.993 billion.

Pre-market quote for Unitree Robotics on Hyperliquid. Source: Hyperliquid

But the sentiment it reflects is real: the market has already priced in an extreme level of optimism about Unitree.

The narrative supporting this valuation is that humanoid robots are about to be widely adopted in factories to replace human labor. But if you look at Unitree’s own prospectus, this story remains just that—a story.

In the first nine months of 2025, 73.6% of Unitree's humanoid robot revenue came from scientific research and education scenarios, 17.4% from commercial exhibitions and performances, and only 9.01% from industrial applications. Of this 9%, less than 3% was generated from actual work scenarios such as manufacturing, inspection, and logistics—amounting to just RMB 15.7 million in the first nine months; the majority of the remainder still came from corporate visits and receptions.

In the article “Humanoid Robots: Before Bringing Benefits, They’re Creating Wealth,” 36Kr raises a sharp question: “If Unitree is profitable, does that mean humanoid robots will definitely be profitable?” and supports this question with specific data.

In the first nine months of 2025, scientific research and education accounted for 73.60% of revenue from humanoid robotics, commercial consumer applications for 17.39%, and industry applications for only 9.01%. Of this 9.01%, corporate tour applications made up 70%, while revenue from clear industrial use cases such as intelligent manufacturing and intelligent inspection amounted to only RMB 15.702 million, representing 29.29% of industry application revenue.

The source of this data is actually Unitree's written response to the second round of inquiries from the Shanghai Stock Exchange:

Yushu Technology's Response Letter. Source: Shanghai Stock Exchange

In short, the largest buyers of humanoid robots today are universities, research institutions, and tech companies, which purchase them for algorithm research, model training, and teaching experiments. However, demand in the research and education market is limited and is likely to reach its ceiling soon.

Of the 9.01% of industry applications, the revenue directly tied to clear industrial scenarios such as intelligent manufacturing, intelligent inspection, and logistics delivery amounts to RMB 15.702 million, accounting for 29.29%. The remaining roughly 70%, according to the company’s own description, is “primarily used for corporate tours.” By “corporate tours,” they mean humanoid robots, after secondary development, performing tasks within enterprises such as greeting guests, guiding routes, providing content explanations, and engaging in interactive Q&A—essentially functioning as talking, walking exhibit displays.

If you multiply the 9.01% industry application share by the approximately 70% enterprise tour share, a significant portion of the revenue counted under the seemingly industrialized "industry application" metric actually comes from front desk reception and exhibition hall tours—far removed from the public’s imagined scenarios of robots assembling products or moving goods in factories.

Revenue directly attributable to intelligent manufacturing, intelligent inspection, and logistics delivery accounts for less than 3% of the company’s total humanoid robot revenue.

The article also notes that while UBTECH’s humanoid robots have secured numerous large orders, they have accumulated losses exceeding RMB 5 billion from 2020 to the first half of 2025 and have yet to turn a profit. As a counterpoint, even impressive order volumes do not guarantee the business is secure. Another analysis article on new listings directly points out that scientific research and education is a profitable business, but it cannot justify the premium valuation attributed to “general-purpose robots entering every household” within the RMB 61 billion market cap. This premium requires support from industrial and consumer markets, both of which are currently still in pilot stages.

In the first quarter of this year, Unitree's revenue reached RMB 423 million, a year-over-year increase of 68.49%. Although this remains robust, it represents a sharp decline compared to the 332.64% growth rate in the first quarter of 2025. Non-GAAP net profit for the quarter dropped from RMB 84.84 million in the same period last year to RMB 40.25 million, a year-over-year decline of 52.55%—effectively halving. The company attributed this to a net increase of RMB 38.33 million in R&D expenses, primarily directed toward embodied intelligence large models, motion control algorithms, and physical structure development, alongside intensified brand promotion efforts on platforms such as the CCTV Spring Festival Gala, which elevated sales expenses.

Financial Response Letter from Unitree Robotics, Source: Shanghai Stock Exchange

Comparing the filing documents disclosed in March with those disclosed in May, Unitree Technology made five revisions to its special risk warnings. The most critical change was the addition of a warning regarding a year-over-year decline in net profit, and replacing the previously broad risk of “failure to achieve technological breakthroughs or product innovation” with “risk of slowed growth and operational performance volatility,” which was moved to the top of the seven special risk warnings. The prospectus also for the first time explicitly named Tesla as a threat, acknowledging that Tesla, leveraging its large-scale production and supply chain integration capabilities, has initiated small-batch trial production of its humanoid robot Optimus Gen-3, with a planned annual capacity of 1 million units; should mass production be achieved, it would directly compete with Unitree.

Wang Feili, an analyst for China’s industrial sector at UBS Securities, noted that even though many manufacturers aim to ship tens of thousands of units this year, humanoid robots may not yet have reached a true commercial inflection point, as current industry orders are still primarily for validation purposes rather than expansion driven by productivity demands.

Zhuo Wang, a partner at an investment firm in Shanghai, told Reuters, "This IPO is overpriced, and the investment risk is already quite high," noting that Unitree's sales are largely derived from research and demonstration scenarios, with large-scale commercial applications still far off.

Even Shenwan Hongyuan Research, which is relatively bullish, used cautious language in its commentary on this pricing, suggesting that the high issuance valuation might draw market attention to the broader commercial potential of the robotics industry, thereby prompting a revaluation of related stocks. Viewed conversely, this also implies that Unitree’s current commercial value remains unproven, with its valuation outpacing its actual business progress.

According to a report by the research firm SemiAnalysis, they estimate that only around 250 Unitree robots will be genuinely deployed in industrial applications in 2025, a significant gap compared to the over 5,500 units shipped that year.

Ultimately, this round of pricing was largely driven by the global surge in interest in humanoid robots and retail investor sentiment in China’s A-share market. Since the beginning of this year, nearly every company associated with the “embodied intelligence” label—from Tesla’s Optimus to Zhiyuan, Yuejiang, and Yunshen—has been granted market valuations far exceeding their current performance.

For this bet to pay off, three assumptions must all materialize simultaneously: a surge in demand for humanoid robots in the short term, Unitree’s ability to maintain market share amid fierce competition from companies like Zhiyuan Robotics and Tesla’s Optimus, and the avoidance of price wars that would erode its current gross margin of around 60%. All three conditions are essential, yet each carries significant uncertainty.

In the short term, due to structural factors such as a small float and index fund support, Yushu’s stock price may not weaken immediately and could even continue to rise in the early stages of listing. However, over the medium term, once these structural supports fade over time—particularly as the one-year anniversary of the listing approaches and institutional lock-up shares become eligible for release—the valuation will ultimately be tested against fundamentals. If revenue from industrial applications does not significantly increase and profit growth continues to decline, this high valuation, currently built on dreams, is likely to experience a substantial correction.

This is so similar to the trend of most crypto assets...