On August 19, 2026, Unitree Robotics (688836.SH) listed on the STAR Market at 150.80 yuan per share, becoming the "first humanoid robot stock" on the A-share market.

On its debut day, the stock price peaked at 1,100 yuan, with total market value briefly reaching 445.5 billion yuan, and the winning bid rate was less than two in ten thousand.

A month and a half later, as of the close on September 30, 2026, Unitree Robotics' stock price had fallen back to 450.40 yuan per share, with a total market value of approximately 182.2 billion yuan, evaporating about 260 billion yuan from its peak, while its dynamic price-to-earnings ratio remained as high as 332.43 times.

Between the surge and the plunge, the market cast a vote with real money: people will buy the technology story, but the commercialization ledger is not so easy to balance.

Technological breakthroughs but lagging commercialization

According to Unitree Robotics' IPO prospectus, its revenue from 2023 to 2025 was 159 million, 393 million, and 1.699 billion yuan, respectively; net profit attributable to shareholders was -11.1451 million yuan, 95.4747 million yuan, and 278 million yuan; and non-GAAP net profit attributable to shareholders in 2025 was 591 million yuan.

On the surface, revenue grew tenfold in two years, and profitability soared.

But when the structure is broken down, problems emerge: in 2025, Unitree's humanoid robot revenue was 868 million yuan, accounting for 51.07% of its main business; quadruped robots contributed 698 million yuan, accounting for 41.05%.

Although the humanoid robot business now accounts for more than half, the robot dog business still occupies nearly half of its revenue.

Moreover, its humanoid robot products are still mostly used for education, scientific research, and other purposes, and large-scale commercial use has not truly begun.

According to The Paper, as of July 2026, Unitree's bipedal humanoid robots had cumulatively rolled off the production line at about 18,000 units, a number that would not even be a fraction of the production capacity of any A-share automaker.

Gan Xiaobin, deputy director of the Science and Technology Department of the Ministry of Industry and Information Technology, publicly stated at the 2026 World Artificial Intelligence Conference that China's full-year humanoid robot production in 2026 is expected to exceed 100,000 units.

The China Academy of Information and Communications Technology's "Humanoid Robot Industry Development White Paper (2025)" also disclosed that global shipments in 2025 were 12,000 units, up 420% year-on-year; the localization rate of core components exceeded 75%, and the localization rate of the three core components—reducers, servos, and controllers—had risen to 75%–90%.

Technology localization is real progress, but there is still a distance between "being able to build" and "being able to make money after building."

Complete machine manufacturers are still losing money

After reviewing the 2026 interim reports of companies in the industry, a counterintuitive phenomenon emerges: complete machine enterprises are generally still burning cash, while component companies have already begun counting real money.

On the complete machine side: UBTech (09880.HK) reported 2026H1 revenue of 1.269 billion yuan, up 104.2% year-on-year; but its net loss was 339 million yuan. Although its full-size humanoid robot revenue surged 1,445% to 590 million yuan, it is essentially still "trading R&D for revenue."

Dobot Technology (02432.HK) reported H1 revenue of 316 million yuan, up 106.6% year-on-year, but net profit attributable to shareholders was -106 million yuan, down 158.68% year-on-year.

Luoshi Robotics (03752.HK) reported H1 revenue of 415 million yuan, up 135.8% year-on-year, but net profit attributable to shareholders showed a loss of 78.97 million yuan.

On the component side: Leader Harmonious Drive Systems (688017.SH) reported 2026H1 revenue of 349 million yuan, up 38.64% year-on-year; net profit attributable to shareholders was 70 million yuan, up 31.25% year-on-year; non-GAAP net profit was 58 million yuan, up 36.92% year-on-year.

Leadshine Technology (002979.SZ) reported H1 revenue of 1.247 billion yuan, up 39.92% year-on-year; net profit attributable to shareholders was 194 million yuan, up 62.79% year-on-year; Q2 surged 92.49% year-on-year. The company has orders in hand for more than 1 million frameless torque motors and is building an automated production line with an annual output of 3 million units.

Inovance Technology (300124.SZ) reported H1 revenue of 24.675 billion yuan, up 20.3% year-on-year, and net profit attributable to shareholders of 2.810 billion yuan.

Beite Technology (603009.SH) reported H1 revenue of 1.204 billion yuan and net profit attributable to shareholders of 63 million yuan, up 14.78% year-on-year.

Wuzhou Xinchun (603667.SH) reported H1 revenue of 1.534 billion yuan and net profit attributable to shareholders of 61 million yuan.

This contrast is very striking: shovel-selling companies like Leadshine Technology and Leader Harmonious Drive Systems are already quietly making money, while complete machine stars like Unitree, UBTech, and Luoshi Robotics burn more fiercely the faster their books run.

According to IIM Information's "Global Humanoid Robot Application Scenario Industry Data and Insight Analysis Report (2026)," the global average selling price of complete humanoid robots has dropped from about $95,000 in 2025 to an estimated $72,000 in 2026. While prices are falling, the cost curve of complete machine manufacturers has not yet been pushed down.

This is a typical "gross margin trap": the high-growth story sounds sexy, but once the cost structure and R&D investment are broken down, the profitability timetable for the complete machine side must be pushed back.

Consumer-grade robots still need time to land

On September 7, 2026, China Daily reported that the industry is shifting from "kung fu mode" to "work mode," with the entire industry entering a "real-scene practical training" stage.

UBS advised investors to avoid homogeneous layouts among complete machine manufacturers and focus on core components such as sensors, actuators, reducers, planetary roller screws, lidar, and tactile systems.

Xu Guangtan of CITIC Securities judged that humanoid robots have entered a "scenario verification inflection point," and core components and industrial scenarios will be the first to scale up.

Yang Zhengye, market director of the National and Local Joint Humanoid Robot Innovation Center, said at the 2026 World Robot Conference that in the next 3–5 years, the easiest commercial value will still be created in industrial manufacturing and service industry scenarios. Household humanoid robots will be difficult to popularize in the short term because the environment is too complex and safety thresholds are high, so there is no need to rush.

Therefore, the priority for humanoid robot commercialization is very clear: industrial inspection > commercial services > special scenarios > household consumption.

Behind this is the dominance of the economic calculus. Customers in industrial scenarios have a strong willingness to pay, an ROI cycle of 2–3 years, the ability to work continuously 24 hours a day, and the avoidance of human fatigue and safety hazards. Commercial service scenarios have short deployment cycles and good display effects, with immediate results.

What about household scenarios? The home environment is unpredictable, flexible operation is difficult, safety standards are stringent, and the cost per unit is high. Household popularization will take at least another 5–10 years.

The first year of commercialization ≠ the first year of earnings realization

So, is 2026 the first year of humanoid robot commercialization? Yes, but only half right.

The first year of commercialization means technology moving from the laboratory to real orders, and robots beginning to "work in factories" in batches.

In the first half of 2026, global humanoid robot shipments were about 19,100 units, with Chinese manufacturers accounting for more than 97% of the share. This number is already enough to excite the industry chain.

But the first year of earnings realization will have to wait until after the component dividend period is released.

Institutions generally expect that 2026–2027 will be the dividend period for core components, and only after 2028 will complete humanoid machines be expected to gradually achieve scale profitability.

For A-share investors, this is a simple multiple-choice question: do not chase the "story premium" of complete machine manufacturers; instead, position for the "earnings certainty" of component manufacturers.

The people who sell shovels always make money first. This saying has been proven for a hundred years in gold rushes, and in the humanoid robot track, it will be no exception.