A week ago, Unitree made its debut on China ’ stock market as the "first humanoid robotics stock." Its IPO price was 150.80 yuan per share, with 40,446,434 shares issued, representing 10 percent of its post-IPO total share capital. The offering price-to-earnings ratio stood at 219.23 times, far above the industry average of 38.56 times.
On its first day of trading, Unitree opened at 1,100 yuan, a 629% increase over the IPO price, sending its market capitalization soaring to 444.9 billion yuan. Founder Wang Xingxing, who owns approximately 121.4 million shares, briefly became the richest person born in the 1990s on paper. The frenzy proved short-lived: by the close of the session the stock had fallen from its opening high of 1,100 yuan to 845 yuan. Subsequent sessions extended the decline, culminating in Wednesday ’ s close at 591.53 yuan and a market value of about 239 billion yuan.
Unitree Lags Behind in Robot Games
At the same time, inside Beijing ’ s National Speed Skating Oval — the "Ice Ribbon" — the second World Humanoid Robot Games were under way. In a preliminary heat of the 100-meter race for large humanoid robots, the Unitree team competed alongside the Tiangong and Lightning teams. Tiangong Ultra finished in 9.39 seconds and Lightning in 9.47 seconds, both faster than Usain Bolt ’ s human world record of 9.58 seconds. Unitree ’ s entry finished in 12.41 seconds — more than three seconds behind the winner and last in its heat.
In the 400-meter final held the same day, Tiangong Ultra won the large-robot category with a time of 38.15 seconds. Unitree failed to win any medals in the sprint events; its only notable result was a silver medal in the Tai Chi final. By contrast, at last year ’ s inaugural Games, Unitree robots won four gold medals, ranking first in both gold medals and total medals. One year later, competitors ’ robots have broken into the nine-second range while Unitree has brought up the rear. Subsequent heats saw times lowered further, with rivals closing the gap faster than expected.
Unitree said on its official Weibo account on the evening of August 22 that, due to constraints on time and manpower, the number of new robots available, and pre-event testing conditions, the team had scaled back several events it had previously registered for. Resources in the past had been focused primarily on mass-produced products. The company also noted that not all of the humanoid robots representing it at this year ’ s Games were operated by its own team; some were developed and entered by customers or partners based on Unitree ’ s platform.
Meager R&D Spending and One-time Customers
Against this backdrop, the market has begun scrutinizing the "first humanoid robotics stock" more closely. Economist Fu Peng told Phoenix Finance that Unitree ’ s valuation before its IPO had already far exceeded the market ’ s broad consensus. After listing, its surge and subsequent volatility were extraordinarily pronounced. The robotics industry remains at a very early stage; although Unitree has completed its IPO, the company is still much more akin to a venture-capital investment at the private-equity stage, he said.
Particular attention has focused on Unitree ’ s research-and-development spending. The prospectus shows R&D expenses of 49.9518 million yuan in 2023, 70.0170 million yuan in 2024 and 145 million yuan in 2025, for a three-year cumulative total of approximately 265 million yuan. The R&D expense ratio fell from 31.39 percent in 2023 to 8.53 percent in 2025 as revenue expanded far faster than spending. By comparison, peers such as UBTECH and DEEP Robotics generally report ratios above 20 percent. Pig-farming company Muyuan Foods spent 1.648 billion yuan on R&D in 2025 — more than 11 times Unitree ’ s outlay that year.
As of the end of January 2026, Unitree held 262 registered patents domestically and internationally, of which only 20 were domestic invention patents. UBTECH, by contrast, had obtained a cumulative roughly 2,800 – 3,000 granted patents by the end of last year, with a majority being invention patents. Unitree plans to invest 2.022 billion yuan of its IPO proceeds in the research and development of intelligent robot models, with a particular focus on the "brain" and "cerebellum." Whether that investment can close the gap remains to be seen.
Unitree ’ s technological strengths are concentrated primarily in motion control — the "cerebellum" layer. Its robots ’ ability to perform backflips, running and jumping is widely recognized. The company is relatively weaker, however, in cognition, decision-making and environmental understanding — the "brain" layer. Its proprietary general-purpose embodied AI model has not yet been deployed at scale in mass-produced products, while industry competition is shifting from hardware specifications toward full-stack capabilities spanning perception, decision-making and execution.
Customer concentration reinforces the early-stage nature of the business. In the first nine months of 2025, 73.6 percent of humanoid-robot revenue came from universities, research institutes and related developers. Commercial and consumer sales accounted for 17.39 percent, while industrial applications contributed only 9.01 percent. Academic and laboratory buyers provide accessible early demand but offer limited scale and weaker evidence of sustained productive deployment.
On the second day after the listing, the company ’ s founder said at the 2026 World Robot Conference that robots still lag behind humans in efficiency and generalization once deployed in factories. Even slight changes in the environment can cause success rates to fall significantly. Embodied AI foundation models remain at an early stage of development worldwide. His timeline for robots' entry into households is two to three years away at the earliest, and perhaps five or even ten years away at the slowest.
The combination of a valuation that briefly priced in rapid industrial adoption, low R&D spending and a research-heavy customer base has driven the swift re-rating of Unitree ’ s shares in its first week as a public company.
