Robot investment: Those who watch the show are the traffic, those who invest money are VCs. Who will bear the losses?
If 2023 marked the beginning of the IPO wave of robotics companies on the Hong Kong stock market, then 2025 is the climax.
On December 29, 2023, the Hong Kong Stock Exchange witnessed an unprecedented event - the bell was rung not by a human, but by a robot.
It was Walker S, manufactured by Ubtech. Standing at 1.7 meters tall, it rang the bell alongside Zhou Jian, the founder of Ubtech. This ceremony was the world's first instance of a humanoid robot participating in the listing bell - ringing event, full of a sense of technology and camera - worthy moments, causing a stir in the capital market.
The scene in 2023 became the opening chapter of the capital fairy tale of embodied intelligence. However, two years later in 2025, the most common question in the media was not whether robots could be manufactured, but who would buy them and whether they could operate sustainably.
When the applause faded, the stock prices fluctuated violently, and institutional investors rushed to reduce their holdings, people suddenly realized a problem: We always say that the ultimate form of a robot is to be like a human. But when such a robot actually appears, what exactly can it do?
In 2025, the Hong Kong stock market's IPO market ushers in the "Robot Season"
In the first half of 2025, the Hong Kong stock market witnessed an unprecedented "robot fever". As of July, a total of three robot - related companies were listed (Geek+, Borayton, and Sanhua Intelligent Control), and 11 robot - related companies submitted listing applications (Standing Robot, SEER, Estun Automation, Woan Robot, Yunji Technology, LeDong Robot, Uisee Tech, Kylas, Yifei Technology, Zhaowei Machinery & Electronics, and Roborock).
Among them, Geek+, the leader in warehouse robots, successfully listed on July 9, raising more than HK$2.7 billion. Its market value at the time of listing exceeded HK$22 billion, but it had a net loss of over 3.5 billion yuan from 2022 - 2024.
Mech - Mind submitted a listing application. It mainly focuses on industrial and medical automation hardware and AI platforms. It had an accumulated loss of 2.281 billion yuan from 2022 - 2024, and its R & D investment accounted for about 40% of its revenue.
Yunji Technology specializes in hotel and commercial delivery robots. In 2024, its revenue was about 245 million yuan, with a gross profit margin of 43%. However, it had an accumulated loss of over 800 million yuan from 2022 - 2024 and faced a gambling agreement pressure of 1.8 billion yuan.
Standing Robot and Kylas Technology also submitted listing applications one after another. They are positioned as suppliers of AGV and AMR platforms. The former had an accumulated net loss of 273 million yuan from 2022 - 2024, and the latter had an annual loss of about 170 million yuan.
Judging from the data, except for a few companies that have made profits on the A - share market and are refinancing through "returning to the Hong Kong market for listing" (such as Fibocom Wireless and Zhaowei Machinery & Electronics), most of the other robot companies are in a long - term loss state. Under Chapter 18C of the Hong Kong Stock Exchange, there are not a few science and technology companies "going public with problems". In other words, this listing boom is more like a capital competition among unprofitable companies: whoever can tell a good story first and raise funds can gain a little more vitality in the brutal industry reshuffle.
In 2023, the Hong Kong Stock Exchange introduced the new listing rule 18C, allowing "specialized technology companies" that have not yet made profits to list. Coupled with the "Technology Company Express Line" launched in 2025, which supports science and technology companies to submit applications confidentially and further speeds up the approval process, it has opened a low - threshold green channel for robot companies. Many robot companies that have not yet made profits but are in urgent need of financing "blood transfusion" have flocked to the Hong Kong market to tell their stories.
The enthusiasm of capital is not only reflected in the number of applications but also in the performance of the secondary market. Benefiting from the artificial intelligence boom and policy dividends, as well as the popularity of the "humanoid robot" concept, the stock prices of listed robot concept stocks have been rising steadily. Since 2025, the stock price of Ubtech, the "first humanoid robot stock", has risen by about 60%, and the stock price of Dobot (which mainly focuses on robot arms and educational robots) listed in Hong Kong at the end of 2024 has soared by more than 170% this year. The Hong Kong market has a relatively high acceptance of emerging technology companies such as robot companies and is willing to give a valuation premium, providing valuable financial support for the development of enterprises.
It's not just the Hong Kong stock market that is hot. The A - share market and the primary market have also been active. Tashizhihang raised the largest angel round of $120 million in March 2025. Unitree announced on July 18 that it had started the guidance for A - share IPO. On July 21, JD.com invested in three robot startups at once: Qianxun Intelligence, Zhujidongli, and Zhongqing Robot, covering the warehousing, logistics, and humanoid robot tracks, almost covering all the current hot topics in the robot industry. Sichuan Tianlian Robot also announced on July 30 that it would go for an A - share IPO.
We can see that although many robot companies are suffering from severe losses and tight cash flows, this does not prevent them from becoming the new favorites of the Hong Kong stock market, with astonishingly high valuations. The confidence of capital in this industry is based on "looking at the future" rather than "looking at the financial statements".
Capital layout and track profile - Who is betting? Who is the most profitable?
These companies that submitted listing applications have attracted a group of established institutions and emerging industrial capital.
A large number of major funds participated in Geek+'s IPO, such as Volcanics Venture Capital, Banyan Capital, Warburg Pincus, CPE Yuanfeng, Granite Asia, and Ant Group, mainly including sovereign wealth funds and other sovereign funds.
Behind Mech - Mind are shareholders such as Sinovation Ventures, AIC, and Joy Capital, including both state - owned capital and market - oriented institutions.
Yunji Technology has received investments from many state - owned and Internet companies such as Tencent, Alibaba, Ctrip, iFlytek, and Lenovo.
Internet giants are scrambling to enter the market. Tencent is betting on the ecosystem, releasing an embodied intelligence open platform and investing in Zhipu AI. Meituan Longzhu Capital has invested in nearly 30 robot companies and has become the second - largest shareholder of Unitree. JD.com has invested in three robot companies such as Zhujidongli at once and incorporated them into its supply chain technology strategy. Alibaba is quietly deploying algorithms and perception systems. While acting as "water sellers" (chips, AI platforms), these giants are also competing for scenarios and hardware, squeezing the living space of small and medium - sized companies.
It can be seen that capital is continuously "staking out territory", and its core logic is: Robots + AI + Industrial manufacturing = Capital blue ocean. Especially for giant platform capital, their investment logic is not only based on future valuations but also on reserving the ability foundation for their own ecological systems.
We can clearly see the different situations of the two major robot tracks:
The first type of robot companies have clear profit paths and stable market demand.
Robotic vacuum cleaners (such as Roborock, Ecovacs, Xiaomi, and Dreame): In 2024, Roborock's net profit was about 1.98 billion yuan, achieving large - scale profitability and building an ecological closed - loop. Dreame had a revenue of about 20 billion yuan and a net profit of about 3 billion yuan in 2024, higher than the industry average. Its high - end strategy has achieved remarkable results - it has a market share of about 30% in the domestic market for robotic vacuum cleaners priced above 5,000 yuan, and the proportion of high - end products overseas exceeds 50%.
Collaborative robots (Cobot): Dobot had a revenue of 374 million yuan. JAKA and AUBO are also actively deploying robotic arms. With mature technology and quantifiable ROI, B - end customers make repeated purchases.
The second type is the "humanoid + platform" track that still relies on financing to survive.
Geek+, Mech - Mind, Yunji Technology, etc.: Although their revenues have increased, their profit models rely on continuous financing to maintain R & D and market expansion. Building their own cloud services, system integration, and robot solution ecosystems still requires substantial financial support.
Humanoid robot companies (Ubtech, Zhujidongli, Zhongqing) are still in the R & D or trial - production stage, and their commercialization paths are still unclear. Their "star" products have sparked global follow - up, but the customer order volume is extremely low, and the market feedback is weak.
In short, the more general capabilities are emphasized, the fewer the scenario matches and orders; the clearer the landing requirements, the easier it is to recover the cost per unit.
The more human - like, the more trapped - Is "humanoid" a step backward?
After Unitree was exposed during the Spring Festival Gala, it showed excellent robot performance. However, although it is the best among all humanoid robots, it only sells tens of thousands of units per year.
New - generation "humanoid - like" companies in China, such as Zhujidongli and Zhongqing Robot, have flexible structures and humanoid appearances but have not yet achieved mass production. Investors are "burning money to bet on the future".
Ubtech's Walker S was once a symbol of domestic humanoid robot technology.
It can walk, open doors, carry objects, dance, go downstairs on its own, and has even appeared on the stage of the CCTV Spring Festival Gala multiple times. In 2023, it "personally rang the gong" on behalf of a Hong Kong - listed company.
It can almost do everything, except being truly marketable.
Ubtech's Walker series of humanoid robots only sold a total of 10 units from 2021 - 2023, with a cumulative revenue of 59.8 million yuan. In 2024, Ubtech also only delivered 10 humanoid robots, with a revenue of about 35 million yuan.
You may ask: What about the dozens of thousands of robots it announced to have sold? They are mainly small - sized products for education, display, or customized scenarios, with low unit prices and thin profit margins. Although they have a certain scale, they cannot support the huge R & D expenses at all.
The problem with Walker is almost the same problem that all humanoid robot companies will encounter: We can make robots more and more like humans, but we can't find useful places for their "human - likeness".
In places where there is real demand - such as nursing, patrolling, security, and logistics - there is actually no interest in "being human - like". Robotic vacuum cleaners don't need legs, robotic arms don't need emotions, and warehouse handling doesn't care if you can dance. It's cheaper for a museum to equip a large number of portable audio guides than to buy a humanoid robot tour guide.
The more human - like, the more expensive; the more expensive, the more useless.
As a result, these robots have become "performance guests" in media exhibition halls and traffic stars in financing PPTs, but they have never really entered production lines, hospitals, or communities.
Ubtech's market value once reached as high as HK$140 billion at the time of listing. Four months after listing, its major shareholder Tencent reduced its holdings by HK$1.034 billion to cash out, and Minyin Capital even cleared its position. To maintain operations, Ubtech had to continuously place new shares for refinancing within a year after its IPO.
Ubtech has failed to make a profit in recent years, with a net loss of 1.124 billion yuan in 2024. Although its revenue has maintained a certain growth in recent years (from 740 million yuan in 2020 to 1.056 billion yuan in 2023), it is constrained by high R & D and market development costs, and the net loss has even increased year by year. From 2020 - 2023, Ubtech's losses were 707 million, 918 million, 987 million, and 1.266 billion yuan respectively, increasing year by year.
You can say that Walker is a great technological product, but you also have to admit that it currently seems more like an over - valued industrial art.
Humanoid design is a way to attract attention, but it's not the logic for selling products. In real - world scenarios, who cares if a robot can shake hands? People are more concerned about whether it can deliver milk tea to hotel rooms, help the elderly get out of bed, or perform medical surgeries accurately - these things are about stable functions, not "pretty appearances".
In contrast, collaborative arms, food - delivery robots, and intelligent conveying systems are more practical. They have gradually entered factories, hospitals, restaurants, and property management, rather than exhibition halls or media events.
Therefore, in the context of "embodied intelligence", humanoid robots have become a topic diverting tool rather than a carrier of mainstream values. Although there are enthusiastic discussions about their visual effects in capital meeting rooms, factory workshops still purchase suitable types of robots based on the previous low - cost and high - stability paths.
The divergence of the "humanoid dream" between China and the US: One is hyping concepts, the other is making acquisitions
China is "believing in the future", while the US is "making acquisitions", which is particularly obvious in the robot industry.
This wave of robot IPOs in China is essentially a "story - telling feast" driven by policies and capital: Local governments have intensively introduced support policies, the Hong Kong stock market has opened the door for loss - making enterprises, the financing channels are smooth, the market is keen on hyping the "ultimate form of general intelligence", and enterprises are competing to tell concepts such as AGI and embodied intelligence.
The US capital market attaches more importance to existing business models: Boston Dynamics was acquired by Hyundai, Kiva was integrated by Amazon, and Universal Robots was acquired by Teradyne. Star companies such as Figure AI have received most of their financing from strategic investments by giants such as Microsoft and Amazon - they rely more on industrial synergy rather than "free competition" in the capital market.
The robot industrial ecosystems in China and the US have different focuses: China is strong in hardware manufacturing and complete supply chains, but still partially relies on imports for high - end core components such as controllers and precision reducers, and the localization process is in progress; the US leads in AI algorithms and software ecosystems, building technological barriers through "giants + venture capital", but has a high external dependence on robot body manufacturing and components.
In terms of the layout of giants, Tesla launched the R & D of Optimus in 2021, reusing the electric vehicle supply chain and Autopilot visual AI, and using the profits from the automotive business to support it. Xiaomi launched humanoid and quadruped robots in 2022, extending from consumer electronics to this field, planning to connect to the smart home ecosystem, and has invested hundreds of millions of yuan in tackling core technologies.
The entry of giants has accelerated the differentiation of the industry: Companies with solid technology and clear scenarios are expected to survive, while small and medium - sized companies relying on concept financing may be eliminated - after all, the endurance of "subsidizing with the money from selling cars" is different from that of "burning VC money". The "dream - building" in the Chinese capital market and the "buying off - the - shelf" in the US will continue to diverge.
After the bubble, the direction will emerge
For companies that have submitted listing applications, the real pressure comes after listing. Robot companies are talking about their visions in the capital market, but after entering the stage of prospectus interpretation and financial statement analysis, institutions and the market are more concerned about profit margins, accounts receivable, customer renewal rates, and product stability. If a company fails to make a profit or reach the break - even point after several quarters of listing, its stock price will quickly be dragged down by the outflow of funds.
Companies such as Ubtech and Zhujidongli that are researching humanoid robots will only be stalling for time to wait for the next financing window if they still cannot achieve mass delivery, no matter how many rounds of financing they have received.
Overall, consumer - grade robots tell us: Make money first, then talk about the platform. Among them, companies such as Roborock and Ecovacs that produce robotic vacuum cleaners have stable consumer groups, clear business chains, and ecological systems.
General collaborative, delivery, and medical robots tell us: As long as orders can be traced and scenarios can form a closed