A Conversation with Yao Haibo from Kunzhong Capital: 2025 is the "survival period" for robot companies. Seizing the initiative is more important than making money.

Author | Tian Zhe

Editor | Su Jianxun

"The application of robots might be quicker than that of autonomous driving, and even surpass it."

When asked about how to view the commercial application of robots, Yao Haibo, the founding partner of Kunzhong Capital, responded.

As an investor, Yao Haibo has been most interested in the hard technology field in the past decade. Shortly after the rise of the autonomous driving industry, Kunzhong Capital participated in the early investment of four autonomous driving companies, and three of them went public: Xiaopeng Motors, which builds the carrier of autonomous driving and landed on the US stock market in 2020; Suteng Juchuang and WeRide, which provide lidar and solutions for autonomous driving cars, successfully went public in 2024.

In 2022, with the wide application of large-scale model technology, Yao Haibo began to pay attention to humanoid robots. After extensive communication with people from various industries, he is more convinced of the value of humanoid robots. Yao Haibo sees that not only nursing homes and factories need robots to work, but even pig farms need robots for cleaning and giving injections.

The imagination space of humanoid robots is vast, but the current technological development is far from perfect. Humanoid robots are still learning simple actions such as folding clothes and pouring water, and there is still a considerable distance from understanding complex human instructions and completing them.

He noticed that many robot companies are facing the same challenge: Selling products is inevitable, but the products still have deficiencies. When is the best time to bring the products to the market?‍

In Yao Haibo's view, in the early stage of the industry, TO B customers' tolerance for products exceeds expectations. They do not have high requirements for the intelligence of robots, but are more concerned about how to improve the operational efficiency of robots.

Although he recognizes the long-term value of humanoid robots, the time left for humanoid robot companies is running out - only one year in 2025. Judging from the perspective of an investor, robot companies either need to obtain a certain amount of income and orders in 2025, or maintain technological leadership in their fields. Otherwise, robot companies have to find their own way out.

This once again reflects the cruelty of the business world. "One of the purposes of investors is to accelerate the maturity of the industry." Yao Haibo said that the robot industry has existed for decades and is an industry that will not close the window for a long time. Although some companies may fail, the industry as a whole will become more and more deeply involved until a key breakthrough or large-scale application occurs.

What is the application path of robot company products?

Yao Haibo's answer is to first develop a special robot for a certain scenario, and then make a general-purpose robot that covers multiple scenarios.

For example, the Da Vinci robot occupies more than 90% of the global medical device market share. Even if it only focuses on the medical surgery scenario, the market value of the "Intuitive Surgical" company behind it can reach 200 billion US dollars, indicating that the market space in the niche field is huge. In the future, it is not excluded that Intuitive Surgical will also develop general-purpose robots.

Recently, around the market application and commercialization problems of humanoid robots, "Intelligent Emergence" had a conversation with Yao Haibo, the founding partner of Kunzhong Capital, trying to observe the current situation of the humanoid robot market from the perspective of investors. The following conversation has been sorted and edited:

Only when the technology is leading or the product can be applied can it attract investors

"Intelligent Emergence": Kunzhong Capital has invested in many autonomous driving companies and achieved some results. Now when investing in robot companies, is there any reusable investment experience?

Yao Haibo: We have adopted a "base area"-style strategy, only focusing on a specific field. We have also created an ecosystem, which consists of three core components.

The first component is lidar, which is the "eye" of autonomous driving; the second component is the "brain", that is, the algorithm of autonomous driving; the third component is the "body", that is, the intelligent car itself.

Based on these core components, we invest in the autonomous driving track. The "eye" Suteng Juchuang and the "brain" WeRide successfully went public in January and August last year respectively, and Xiaopeng Motors equipped with Suteng lidar has achieved a significant leading advantage in the intelligent driving solution.

From the investment layout logic of the previous cycle, if we want to invest in an embodied intelligent device, we will first cover the core components, then the algorithm level, and finally the product itself.

Regarding robot investment, we also plan to follow a similar layout, specifically including the three major capabilities of movement, operation, and perception.

"Intelligent Emergence": Now robot companies have their own advantages, but in terms of commercialization and orders, they are basically at the same starting line. When you invest, how do you find the potential of the company?

Yao Haibo: All investment funds are focusing on how to improve the hit rate. Kunzhong Capital pursues the hit rate with several key premises.

The first premise is to enter the game early. Just like Kunzhong Capital's investment in autonomous driving, if we laid out the autonomous driving in 2014, according to our experience, we would basically be able to successfully find one or two leading companies. But if you start to lay out in 2016 or 2017, then you may not even have the opportunity to see the leading companies because they are already in rapid development.

The second premise is to find the best entrepreneurs. Excellent entrepreneurs usually have a very strong academic background, especially in the field of robotics.

For example, the CEO of Zhujidiandong that we invested in, Teacher Zhang Wei, collaborated with the founder of Boston Dynamics when he was studying and teaching in the United States. And the co-founder and chief scientist of the Daimeng Robot that we invested in, Teacher Wang Yu, founded the Robotics Research Institute of Hong Kong University of Science and Technology as the founding dean.

"Intelligent Emergence": Kunzhong Capital also invested in Fengtan Robot, a company founded by an industry veteran. When investing in such a company, which elements will you particularly value?

Yao Haibo: Fengtan is an attempt of ours. In the application of robots, we have not yet fully determined the future application scenarios and still need to explore.

The most realistic problem of application exploration is commercialization. We will pay special attention to talents with vertical field sales experience and cross-border sales capabilities.
 

For example, the Fengtan team has rich overseas sales experience. Its CEO, Li Zike, served as a senior executive of a listed company before starting a business. He has been engaged in the research and development and promotion of construction technology products for decades, and many products have become industry leaders. This is a great attraction for us.

"Intelligent Emergence": What aspects of robot companies does Kunzhong Capital currently focus on?

Yao Haibo: For the companies we focus on, they must first have a certain technological leadership and a clear product application. Now, except for the consensus in the field of humanoid robots, there are no unified standards and consensus in other aspects, and many attempts in technology and hardware are still in the exploration stage.

The biggest confusion at this stage is that before the robot is finalized, many basic models of hardware may be made, but they cannot really solve practical problems. Therefore, we value more companies that have already made clear breakthroughs in application scenarios. For example, if a company can come up with a world-leading all-terrain autonomous mobile solution, whether it is a four-wheeled, two-wheeled or humanoid, this innovation is of great value.

Selling products is more important than making good products

"Intelligent Emergence": Recently, I communicated with some industry insiders. They mentioned that if the company cannot generate income this year, it will be difficult to obtain financing in the future and may be eliminated by the market in 2026. How do you view this industry sentiment?

Yao Haibo: I think this is a very correct understanding. First of all, we should indeed thank those customers who are willing to pay at present, because most of them are still in the trial stage. After all, there is no robot that can completely break the market and bring a clear input-output ratio to customers.

However, the current situation of the robot industry is that even if your product is the best in the industry in technology, it may not be able to win the market or make money immediately. This is an inevitable stage in the field of embodied robots, because the early development of many innovative fields cannot directly bring profits through quality.

Many robot companies that we invest in do not particularly emphasize the profit requirement in this year's goals, but require them to be "fast".

Fast means that you have to be faster than others in technology investment, technology exploration and application exploration, and remain at the forefront of the industry.

Of course, being fast does not mean not making money. What we mean is that your core task is to maintain leadership. And, I summed up a sentence: "Brave climbing to the peak and laying eggs along the way".

What is brave climbing to the peak? It is to be the top in the industry, which is the primary goal.

But at the same time, laying eggs along the way is also very important. Even if your product may not be perfect in quality and there are some problems, such as repair or the failure rate of operation in the real scene, you still need to sell it and enter the customer supply chain. If you don't enter now, in the future, when the customer's technology is successful, you may miss the opportunity to be a supplier.

The combination of the two goals of technological leadership and the ability of market application is the ability that robot startup companies should possess most at present.

"Intelligent Emergence": Autonomous driving has been developing for nearly ten years since its big outbreak in China in 2015. However, in terms of commercial application, especially Robotaxi, only a few companies have achieved operation without safety officers, and the annual income of these companies is only at the level of several hundred million. Relatively speaking, the embodied robot industry broke out in 2022 and has only been more than two years so far, but the commercialization pressure of such companies is particularly high. Are investors accelerating the maturity of the robot industry?

Yao Haibo: Investors are the catalysts of revolutionary and innovative technologies, helping startup companies succeed quickly.

Indeed, investors have promoted the rapid development of many fields, but if you ask me which industry is more likely to achieve industrialization or make profits faster, I think the robot industry may take the lead in making profits.

Although I have also heavily invested in the field of autonomous driving, and autonomous driving companies have already had a certain scale of income, but I still think that the application of robots may run faster than autonomous driving, and even surpass it.

The reason is that the business model and benefits of robot applications are relatively clear. For example, Tesla announced this year that it will mass-produce 10,000 Optimus, which is a considerable goal, indicating that they want to accelerate mass production. These mass-produced robots are relatively easy to achieve, but the automotive industry has higher limitations and uncertainties due to the need to interact with the external environment.

And the robot industry has many application scenarios in known environments, which can achieve income more quickly. For example, cleaning and injection robots in nursing homes, factories, and even pig farms, these scenarios are already visible and tangible and can be commercialized quickly. Moreover, the return on investment in these industries is relatively direct and can bring visible benefits.

I don't think we need to worry about not seeing income in the short term. Although this industry requires a huge investment and investors are catalyzing this process, there are still many practical expectations and opportunities.

"Intelligent Emergence": Autonomous driving has a mature carrier like a car, while the hardware cost of the robot industry is still relatively high, and robots are still in the stage of just learning to walk or run. When the development of the robot itself is not yet fully mature, pushing them to the market may make customers feel that the products are not good enough, which may indirectly affect the progress of the entire industry?

Yao Haibo: Now all investment enterprises and leading companies in the industry are facing this challenge. First, you have to ask yourself whether you want to sell or not? Secondly, is the product you sell good enough?

I think the answer to the first question is definitely to sell, and the answer to the second question, whether the product is good enough, is "not good enough". As you said, these challenges exist, but I think the tolerance of customers for these products is higher than we expected.

Now customers' requirements for robots are not 100% perfect. For example, Ruoyu Technology has a refueling robot. The customer demand in this refueling scenario is not high. It only needs the robot to move from point A to point B, insert the fuel gun, and then pull it out. In fact, the customer's requirement for the intelligence of the robot is not high, but more to solve the problem of improving efficiency and operability. Customers will tolerate some small flaws in the robot's execution of the task as long as it can complete the task.

Therefore, I think this is a particularly good time. If you keep "training troops" at home, thinking of making a perfect 100-point product before selling it, you may miss this wave of the earliest experience users. These customers are buying robots now not for a pure return on investment, but more for paying for the experience and attempt.

In addition, we can review the success of the past few generations of robots, and they are all related to a certain industrial dividend. For example, the earliest logistics robot developed with the dividend of cross-border e-commerce, and later the self-driving forklift in the factory is related to the dividend of the expansion of new energy and lithium batteries. Now, everyone is looking forward to the next industrial dividend. Although these early investors and users have not yet seen huge returns, they are willing to try and pay for it, which is itself an investment opportunity.

So, if you don't dare to try to apply now and miss this point, when the next dividend comes, you may miss the opportunity to enter this industry. Regarding the trial period, I think it is about one to two years, because there is still demand at this stage. Even if the product is not perfect, customers are willing to pay for this new experience. Moreover, the government's policy support and subsidies also provide an opportunity window for these early applications.

"Intelligent Emergence": The entire robot industry and Kunzhong Capital are looking for good scenarios for the commercial application of robots. You mentioned that robot companies need to be given enough patience. Has Kunzhong Capital set a clear commercialization deadline for the invested companies?

Yao Haibo: We do not have a particularly clear commercialization indicator or KPI, especially in the commercialization direction. Of course, we will provide help, such as helping them find a suitable commercialization partner to jointly promote industrialization.

For example, we have done a lot of work in the post-investment of Zhujidiandong, helping to introduce a co-founder and join their team to promote industrialization. We are not just providing funds, but will build together with the company.

I think the most important thing is to help them find the right landing partner, rather than just setting a KPI and asking them to execute according to our standards.

"Intelligent Emergence": If there is no obvious KPI, how much annual income can be considered as reaching the standard or being successful?

Yao Haibo: From an investment perspective, we do not simply rely on income as a measurement standard. We will make budgets and final accounts with the company every year. In terms of commercialization, we pay more attention to whether it can become the industry leader. For example, technological leadership is one aspect, but more importantly, whether it can stand out in the industry and become a leader.

For example, Suteng Juchuang has now achieved the world's first shipment of lidar in terms of production capacity and sales. We hope that Suteng Juchuang can win the top customers, but this requires time. Therefore, we will not simply set a KPI based on whether we can sign several major customers within one year. What we value more is whether we can win a top customer like BYD, even if it is a single major customer.

"Intelligent Emergence": If in 2025, robot companies on the market have not obtained significant income or orders, what will happen to these companies in the future and what ways can they survive?

Yao Haibo: If a company has no technological leadership, no top customers, and the team is not the top group, I really think its survival will face huge challenges. Some people say that 2025 is the order year, and some people say it is the application year, but I think this year may be the "survival year", and the fittest will survive. It is not that most companies are hopeless, but at least you have to have a "Chinese record or world record" in one aspect to get more attention and support in the market.