Interview with Early Investors of Unitree and Rokid: In Hangzhou's "Seven Dragon Balls", We've Successfully Backed Two of Them | Hardcore Kr Interview
Author | Zhang Ziyi
Editor | Yuan Silai
Yu Wenchao, an investor from Dunhong Asset, has been asked more than once how he managed to invest in Unitree in 2022.
Yu Wenchao also ponders over this question. He is not so arrogant as to attribute it solely to his judgment - though it is important - but luck also played an essential role in Unitree's rise.
Anyway, when Yu Wenchao first met Wang Xingxing, he couldn't have imagined that this lean young man in front of him would sit in the same row as Ren Zhengfei and Wang Chuanfu four years later, becoming a kind of benchmark and symbol.
Unitree's old shares are now one of the hottest and most expensive acquisition targets in the primary market. Investors privately know that the company is not far from going public. If someone gets a ticket, they are likely to make a profit.
Back in the autumn of 2021, when Yu Wenchao decided to invest in Unitree, it was only moderately popular at best and far from being a sought - after target.
As a founding partner of Dunhong, Yu Wenchao has been involved in a series of star projects, including the listed companies XGIMI and ArcSoft Technology, and the unicorn Dangbei. He has also participated in companies with excellent technology but poor sales. However, these achievements don't make for a very dramatic story, so outside the circle, Yu Wenchao and Dunhong Asset are not well - known.
Unitree was also rather obscure before 2025. During the Spring Festival Gala of the Year of the Ox, Unitree's quadruped robot dog appeared on the stage, and Yu Wenchao began to notice the company. However, the brief dance of the robot dog with Andy Lau was soon forgotten by people locked at home.
The most feverish investments that year were in the consumer goods sector: Momo Pastry Bureau raised funds in three rounds in half a year with only 14 stores, and the value of one store exceeded 100 million yuan; after a new round of financing of $200 million, the valuation of Genki Forest climbed to $15 billion, higher than Mengniu, which has been established for more than 20 years.
Besides consumer goods, "Metaverse" and Web 3.0 took up the rest of the topics, and rounds of investments were like a game of hot potato. As for the robotics sector where Unitree is located, investors couldn't see any exciting business prospects. At best, this sector was just mediocre.
Yu Wenchao never thought of investing in the catering industry. "We are a group of science and engineering guys and really don't understand milk tea." Since 2018, Dunhong Asset has mainly focused on hard - tech. Their investment team is from Zhejiang University, Shanghai Jiao Tong University, and Shanghai University of Finance and Economics. Their investment style can be described as practical and steady, definitely not radical.
Of course, he will also follow the trend boldly in industries he is familiar with. For example, after missing out on PICO, he quickly invested in Rokid - this AR glasses company has a founder with the visionary spirit of Silicon Valley. It has been established for more than a decade, and its profitability has always been a mystery, but it has become a shining star in Hangzhou alongside Unitree, and you need to make an appointment in advance to visit its exhibition hall.
More often than not, compared with grand and illusory stories, the financial situation is the key indicator for Yu Wenchao to evaluate a project. When Yu Wenchao first met Wang Xingxing in Hangzhou, he "didn't have a deep understanding" of embodied intelligence. However, what really made him place a bet was that Unitree had crossed the survival line, was doing well in the B - to - B business, and already had C - to - C products. "Unitree Technology has a relatively high lower limit, and it's worth betting on its upper - limit development potential," he recalled.
When deciding to invest, Yu Wenchao expected Unitree to become the "DJI on the ground". As the investment was finalized, he felt that Unitree had given them more pleasant surprises and updates than expected.
Wang Xingxing has the unique paranoia of a tech enthusiast, and at the same time, he is flexible and adaptable like a businessman. This may be related to the fact that Wang Xingxing is from Ningbo, Zhejiang, a city with a highly active business culture. Wang Xingxing has a natural sensitivity to making money.
Yu Wenchao asked Wang Xingxing more than once: "Will you make humanoid robots?" The answer was very straightforward: "No." But later, when the technology matured and customers came to him again and again to place orders, Wang Xingxing jumped in without hesitation. Two years later, he was able to deliver dozens of robots - this mass - production ability is extremely rare among humanoid robot startups.
Moreover, Wang Xingxing is well aware that he needs support beyond just money, including supply chain resources and connections with leading media.
These are exactly what Dunhong Asset has. Their executive director graduated from Nanjing University and has good relationships in Zhejiang, Shanghai, and Hunan. He accompanied Mango Excellent Media (the parent company of Mango TV) to go public and also holds a position in an official institution - these words represent local, down - to - earth, and in - system resources.
After all, Unitree doesn't have much time to spare. In the humanoid robot sector, large companies like Huawei and Xiaomi haven't entered yet. Once the giants get involved, small and medium - sized companies in the industry will find it difficult to resist. This has been proven in capital - intensive industries such as smartphones and new energy vehicles.
At this uneasy moment, Unitree has gathered a lot of emotions. The primary market regrets missing out on Unitree and Rokid and is also troubled by the transformation after the bonus fades; entrepreneurs are confused about the direction of entrepreneurship, and even star companies like Unitree have their own difficulties. We had a chat with Yu Wenchao from Dunhong Asset, hoping to find some solutions and gain some confidence.
01 About the Investment in Unitree Technology
Yingke: How did you first get in touch with Unitree Technology, and what was its development status at that time?
Yu Wenchao: I first noticed it during the Spring Festival Gala of the Year of the Ox in early 2021. Their product was painted like a little calf and performed on stage with Andy Lau. Since they are in Hangzhou, we went to have a communication. However, it was a bit late. At that time, their round of financing was almost completed, and we missed the opportunity. But we believe that Unitree Technology is a new thing with great potential, and the team is very good. When a new round of financing started in the second half of 2021, we quickly followed up and were the first institution to pass the investment decision in that round.
Yingke: What were the reasons for investing in Unitree Technology at that time?
Yu Wenchao: Before investing in Unitree, we had looked at many robots, such as the popular collaborative robots at that time. But different from other industrial robots and collaborative robots, we actually didn't regard Unitree as a robot project but as an intelligent terminal. Unitree's robot dog is an innovative "new species" with the potential to enter the household consumer market.
In 2021, it launched its first - generation C - to - C product, Go1, with a price of about 12,000 yuan. Although its functions were relatively simple, with only basic accompanying functions, it was the first consumer - grade robot product globally, which was of pioneering significance. We judged that as the product functions become more and more powerful and the price drops to the thousand - yuan level, it has a good chance in the consumer - grade market and may even become the "DJI on the ground". Before drones added aerial photography functions, the market demand was very small. After adding various aerial photography functions, the market suddenly opened up. We think the Go series products will be the same.
There are two other important points that support our investment in Unitree. The product technology and innovation capabilities of the Unitree team are extremely strong. From the time we started contacting them in early 2021 to the completion of the investment in late 2021, we could see new surprises almost every one or two months. After the investment, every two or three months, the product innovation, iteration, and performance always exceeded our expectations, which left a very deep impression on us. At that time, humanoid robots were not that popular, and our expectation was the "DJI on the ground". We didn't expect that the company's underlying technology and product innovation capabilities would fully support the current humanoid robot story.
Another important point is that the company's lower limit is at least guaranteed. At that time, Unitree Technology's main revenue business was concentrated in the B - to - B field. We believed that with the team's strong product technology capabilities, at worst, it could reach the level of a listed company on the Science and Technology Innovation Board that makes inspection robots. That company started with the inspection business of the Jiangsu Power Grid in its early days. At its peak, its revenue reached more than one billion yuan, its net profit was tens of millions of yuan, and its market value once reached tens of billions of yuan. So we thought Unitree Technology had a relatively high lower limit, and it was worth betting on its upper - limit development potential.
Yingke: Were there many competing institutions when you invested in Unitree Technology? Why were you able to become the co - leading investor?
Yu Wenchao: After all, Unitree had appeared on the Spring Festival Gala. Although it wasn't as popular as it is now, it was still a hot project in the industry, and basically all institutions would take a look. The competition for the quota in that round was also very fierce. Well - known institutions such as Shenzhen Capital Group and China Internet Investment Fund, as well as Hexagon, a Fortune Global 500 company acting as the business coordinator, were all follow - on investors.
The key reason why we were able to co - lead the investment with Matrix Partners was that Unitree Technology had launched its first - generation Go1 product at that time and planned to carry out C - to - C business. And we had invested in TMT in the past and had strong Internet traffic resources. Our LPs included Zhejiang Satellite TV and Mango TV, so we had very rich media resources, which could support Unitree Technology's C - to - C business. In addition, we had previously invested in well - known intelligent terminal companies such as XGIMI and Dangbei, and had accumulated resources such as chip supply and e - commerce, which could help Unitree Technology make connections. At that time, Unitree Technology's revenue was in the tens of millions of yuan, and the chip manufacturers might not pay attention to them directly, but we could find the sales managers or senior executives of IC companies for direct communication, which was a certain empowerment for the enterprise.
Moreover, we were also the first institution to pass the investment decision.
Yingke: What's your impression of Wang Xingxing? I heard that he is not very good at expressing himself.
Yu Wenchao: Wang Xingxing is a very pure science and engineering guy. There are very few such founders in China. When communicating with him, you can feel that his mind works very fast. He fully understands the technology, and his speaking speed is also very fast, so the communication efficiency is high. And he is from Ningbo, Zhejiang, where the business atmosphere is very strong, so his business sense is very sharp.
From the team perspective, Unitree Technology established a complete business team at the beginning of its establishment. Chen Li, the co - founder, is Wang Xingxing's classmate and used to work at Hikvision. He has strong B - to - B and B - to - G sales capabilities and rich experience in this area, so he is very suitable for the sales work.
The company has formed a leadership structure with the CEO also serving as the CTO in the first position and a co - founder with a BD sales background in the second position. The whole team can start from customer needs and is sensitive to market changes.
Wang Xingxing said in an interview before that a few years ago, an investor asked him if he wanted to make humanoid robots, and he refused without hesitation. "Whether to make humanoid robots" was the question I asked him that year. It was still early at that time. Tesla's Optimus hadn't come out yet. Elon Musk was like issuing a call to the world, saying that Tesla was going to do something in the humanoid robot field and was recruiting heroes to join.
Wang Xingxing was very firm at that time. His exact words were that he thought bipedal robots had no commercial value and no landing scenarios had been found. I think this judgment was correct. After all, we didn't have a deep understanding of humanoid robots at that time.
In 2023, when customers asked about the humanoid robot business, they thought there was market demand and it wasn't difficult to do, so they immediately adjusted their strategy according to the market changes.
Yingke: What do you think of the future development of humanoid robots? Do you have any predictions for Unitree Technology's business in this area?
Yu Wenchao: I believe that China will lead the world in the embodied intelligence/humanoid robot industry in the future. There are three reasons. First, in the field of general language large models, China is restricted by computing power, but in the field of embodied intelligence, the demand for computing power is not that high. Currently, thousands of computing cards are enough, and China's computing power reserves are sufficient to meet the R & D needs of embodied intelligence. Second, in terms of robot hardware, China leads the world with its advantages in three - electric technology and manufacturing supply chain. Third, China also has advantages in downstream application scenarios and data.
The main landing scenarios of embodied intelligence in the next few years will be in the manufacturing industry. China has a large number of manufacturing enterprises, a large number of application scenarios, and many large enterprises are very willing to spend money (about the salary of a worker for two years, 200,000 - 400,000 yuan) to buy products that meet their needs. In terms of landing scenarios and application data, China's embodied intelligence/humanoid robot companies have more advantages and are more likely to achieve commercialization.
Currently, the humanoid robot industry is still in its early stage of development, and the biggest bottleneck lies in the robot's brain. The brain part (embodied large model) is still in continuous iteration, similar to the GPT 1.0 and GPT 2.0 stages before GPT 3.0. Although it can already be applied, it is not powerful enough, but the iteration speed is very fast. In terms of hardware, if it wants to enter the consumer market, the price needs to be further reduced, and Unitree Technology has already made a breakthrough in this regard.
In the field of humanoid robots, Unitree has reached the world - class level in terms of the robot's body and cerebellum. I even think that they have advanced the industry by two or three years on their own, and they have "beaten" the price to many people's psychological expectations.
Previously, the mainstream embodied intelligence humanoid robots in the market were for factory applications. Factories calculated how many employees a robot could replace, and they could accept a price between 200,000 and 450,000 yuan. However, Unitree Technology's price of 99,000 yuan not only makes it acceptable to many small B - customers but also attracts some geeky C - customers to buy and experience.
02 About the Investment in Rokid
Yingke: Why did you choose to invest in Rokid?
Yu Wenchao: Before investing in Rokid, we had actually been looking at the XR industry for a long time. We first looked at PICO.
We started talking to PICO in the summer of 2020. They offered us a very low price at that time, and we thought it was good. VR was popular in 2015 and 2016 and then declined. The market expectations were too high. After 2020, VR technology had improved after iteration, and PICO was one of the better - performing companies that had survived. We were quite complementary to PICO. They needed content, and we had resources of Internet content platforms. Later, after the due diligence was completed, there were some objective reasons to wait before investing. As a result, in early 2021, when Oculus was launched and became very popular, PICO also took off, and many institutions rushed in. By the time we wanted to invest, it was a bit too late.
After missing out on PICO, we also made a judgment: in 2021, although AR technology was not yet mature, in the long run, its application scenarios were more extensive. The future development prospects of AR were better than those of VR.
VR is mainly used in immersive entertainment scenarios and is restricted by the environment. It is usually used in specific environments such as at home and is more like a game console. In contrast, AR products are more portable and can be widely used in various scenarios such as work, entertainment, and learning.
One of our investment strategies is to only invest in the leading companies in niche markets. If we can invest in the