The so-called "first humanoid robot stock" Unitree Technology made its STAR Market debut on August 19, opening at 1,100 yuan against an issue price of 150.80 yuan—a staggering 629.44% surge that briefly pushed its market value to 444.9 billion yuan. What followed was a relentless freefall. The stock closed its first day at 845 yuan, trimming the valuation to 341.8 billion yuan, then slid 18.70% on day two, 2.12% on day three, and a further 10.31% on the fourth trading day to settle at 603.08 yuan, bringing the total market cap down to 243.9 billion yuan.
From its intraday peak on listing day, the cumulative drawdown has now exceeded 45%, erasing roughly 200 billion yuan in market value. Retail investors are venting frustration over what they see as an egregiously inflated valuation, with one commenter quipping that the frenzy leaves nothing but scraps for secondary market participants. The pressing question remains: what is Unitree actually worth, and how far must the stock fall before it becomes a reasonable entry point?
Answering this requires two different yardsticks—one for the secondary market, and one for the private equity stage. Once you measure with both, it becomes clear that the ones losing sleep should not be retail traders.
The 444.9 Billion Peak Was a Liquidity Mirage, Not a Valuation
The 444.9 billion yuan peak on day one was never a genuine secondary market price discovery. Unitree has a total share count of 404 million, but only 30.09 million shares—a mere 7.44% of the total—were freely tradable at listing. With no price limit for the first five trading days and scarce float amplifying speculative fervor, a relatively small amount of capital could push the price to an absurd level. It is like pricing an entire cold storage warehouse based on the transaction price of just one displayed basket at a vegetable market.
Capital flows confirm this distortion. The first day saw turnover hit 85.28%, with trading volume reaching 23.16 billion yuan, as free float changed hands completely within a single session. Super-large orders saw a net inflow of 4.42 billion yuan, while large and medium orders posted net outflows of 2.22 billion and 2.19 billion yuan respectively—both ranking first across the entire A-share market. In other words, hot money and retail investors were buying while institutions and IPO profit-takers were handing over their chips. The 444.9 billion peak was a "free-float illusion" pricing, generated by leveraged sentiment on just 7.44% of shares. Using it as the starting point for valuation discussions only leads to nonsensical conclusions.
The Premium Is Not Just Slightly Excessive
On the secondary market, Unitree's first meaningful price point was its IPO pricing at roughly 60 billion yuan, a level most institutions found acceptable. That number is not entirely unreasonable. However, at its current stage, Unitree remains fundamentally a manufacturing company. In the first nine months of 2025, humanoid robot revenue reached 595 million yuan, with 73.6% coming from scientific research and education clients—universities and research institutes. Commercial consumption scenarios contributed 17.39% of robot revenue, while industrial applications like corporate tours, smart manufacturing, and intelligent inspection accounted for just 9.01%. Today's buyers are purchasing Unitree robots for teaching, R&D, and demonstration, not to replace workers and generate cash flow. These are still premium toys, not yet productivity tools.
As a manufacturing entity, Unitree should be measured against manufacturing sector benchmarks. Comparing it to six representative listed manufacturers—including a power battery leader, a home appliance giant, an industrial control champion, and fellow robotics firms—reveals a stark picture. A-shares already have a clear valuation ceiling for manufacturing. CATL, the global leader in power batteries with a generational technology lead and high growth, trades at a static P/E of roughly 25 times. Inovance Technology, the sharpest tool in domestic substitution, commands 31.3 times—already on the higher end for manufacturing. Unitree's static P/E stands at 877.4 times, far exceeding all six peers.
Market capitalization paints an even more vivid picture. Unitree's 243.9 billion yuan value equals roughly 7.6 times Ecovacs, yet Ecovacs generates 19 billion yuan in annual revenue and 1.76 billion yuan in net profit—11 times and 3 times Unitree's figures respectively. Meanwhile, Hikvision, with annual revenue of 92.5 billion yuan and net profit of 14.2 billion yuan, holds a market cap of 312.9 billion yuan. Unitree's valuation is nearly 80% of Hikvision's, but its revenue is just 1.8% of Hikvision's. A P/E of 877.4 times and a P/S of 144 times is not a matter of being slightly expensive—it is completely disconnected from the entire reference framework.
Even granting Unitree a premium of twice Inovance's multiple—60 to 80 times P/E—and optimistically projecting 2026 net profit to reach 700 million yuan, the implied market cap would only be 42 to 56 billion yuan, roughly aligning with the IPO price. The narratives that could justify a higher valuation hinge on Unitree's growth trajectory, its "brain" capabilities, and liquidity premiums. While robotics is indeed a growth sector, the certainty of expansion is not ironclad. Industry involution and intensifying competition could compress gross margins. On the AI brain front, Wang Xingxing noted at the World Robot Conference that the "ChatGPT moment" for humanoid robots may still be 2 to 10 years away.
Although Unitree is already profitable, some institutions are spinning more expensive stories using sales multiples. CCB International applies a 32 times target P/S for 2026, yielding an approximate valuation of 109 billion yuan. Nomura has issued a "Buy" rating with a target price of 370 yuan, corresponding to a market cap around 150 billion yuan. Given current liquidity premiums, a valuation above 100 billion yuan remains barely defensible, but anything exceeding 200 billion seems excessive without substantial brain-model breakthroughs. The key to unlocking a higher valuation is Unitree's ability to leverage large models to enter more scenarios—a breakthrough that needs to materialize within two years but remains uncertain, thereby discounting the value attributed to its AI capabilities. The reasonable valuation range for Unitree lies between 60 billion and 150 billion yuan, with a midpoint around 100 billion yuan, corresponding to a share price near 250 yuan. The 150 billion mark represents the ceiling if all narratives align perfectly. At 243.9 billion yuan today, the stock remains over 60% above the optimistic anchor—the bubble is half squeezed, but not yet fully deflated.
Unitree Is the Pricing Anchor for the Entire Sector
Turning to the primary market, the math gets even more intriguing. In June 2025, Unitree's final private funding round valued the company at 12.7 billion yuan post-investment. Against projected 2025 revenue of 1.7 billion yuan, that implies a P/S of about 7.5 times; against 270 million yuan in net profit, a P/E of roughly 47 times. For a high-growth enterprise, that is not unreasonable.
The real problem is that the entire primary market now prices humanoid robot companies not against Unitree's 12.7 billion private valuation, but against its post-listing market cap of 300 to 400 billion yuan. As of June this year, eight domestic embodied intelligence companies have reached 20 billion yuan valuations, forming a "20 billion club"—including Agibot, Galbot, Xinghaitu, Qianxun Intelligent, Zibian, Zhipingfang, and Lingxin Qiaoshou. Many of these companies still ship in the hundreds of units, yet their funding velocity is aggressive: Galbot has raised approximately 7 billion yuan cumulatively in under three years, Zhipingfang completed 12 funding rounds in one year, and Zibian closed four rounds in two months. A company shipping just 100 units being valued at 20 billion yuan implicitly assumes that Unitree—shipping 5,500 units with a 32.4% global market share—is worth 200 to 300 billion yuan. Unitree's secondary market cap serves as the "denominator" for the entire primary market's valuation system. Every drop in Unitree's share price shrinks that denominator, forcing primary market valuations to be renegotiated. If Unitree stabilizes near 100 billion yuan, companies with revenue a fraction of Unitree's but valuations 1.6 times its private level will find investors refusing to sign off on new funding rounds. Many such companies will struggle to raise capital—this is the deepest implication of the current decline.
According to media reports, an estimated 30 to 50 robotics companies are preparing for Hong Kong listings. The vast majority have yet to prove their business models, yet their valuations have already soared to the heavens, with IPO as their only exit route. That exit channel's value is depreciating in tandem with Unitree's share price. Notably, funds that entered between 2021 and 2022 are now entering their exit windows, and many project financing agreements include repurchase clauses. On the demand side, Goldman Sachs projects global humanoid robot shipments will reach just 76,000 units by 2027, rising to 502,000 by 2032. The sector is real, but the pace of delivery cannot keep up with the speed of valuation inflation. Consequently, the primary market does not want to see Unitree's valuation fall below 200 billion yuan—let alone below 100 billion yuan.
IPOs also operate within policy windows. Unitree's listing journey from acceptance of its application on March 20 to its August 19 debut took just 152 days—a record under the STAR Market's "pre-review" mechanism. Regulators opened the door quickly and widely, but such windows do not stay open forever. The next 6 to 12 months represent the "escape window" for embodied intelligence primary market players. Companies that complete IPOs or M&A exits before Unitree's valuation anchor sinks further can still ride the wave with inflated valuations. Those that delay will face the brutal reality of primary valuations converging downward to secondary levels. First movers win; latecomers suffer.
That said, the primary market is not entirely without life support. Three conditions could reverse the logic: first, a substantive breakthrough in the generalization capabilities of embodied large models; second, genuine large-scale industrial orders rather than repeated purchases of research and teaching equipment; and third, Unitree delivering several consecutive quarters of earnings that exceed expectations, gradually digesting its valuation. Currently, none of these conditions has materialized.
The Bottom Line
Where does Unitree's decline finally end? In the secondary market, the reasonable band is 60 billion to 150 billion yuan, with a midpoint of 100 billion yuan corresponding to a share price around 250 yuan. At the current price of 603.08 yuan and a market cap of 243.9 billion yuan, there remains significant distance to the upper bound of that range. Secondary market investors should monitor four key signals: the pace of lock-up expirations (with only 7.44% of shares freely tradable, unlocking represents a sword hanging overhead), quarterly changes in the share of industrial revenue, gross margin trends as average robot prices decline, and whether embodied large models achieve tangible generalization progress. These factors will all influence Unitree's valuation.
For the primary market, Unitree's floor is their ceiling. Every increment of the bottom being probed narrows the window further. This 440-billion-yuan fireworks display illuminates not just Unitree's share price, but also the countdown timer facing embodied intelligence companies in the primary market.