From a Garage to 42 Billion: The Hard-Won Rise of Unitree Robotics and Its Mysterious Backer
On June 1st, Unitree Technology, which has been established for ten years, passed the review, becoming the company that passed the review the fastest since this year: it only took 73 days from submitting the materials to passing the review. It is estimated that the issuance market value will be 42 billion yuan. The industry expects that Unitree will complete all the procedures in July this year and win the title of "the first stock of embodied intelligence/humanoid robots in A-share".
At that time, Unitree had just completed its first decade. In August 2016, Wang Xingxing founded Unitree Technology in an unremarkable office in Hangzhou.
For this small-town youth born in Yuyao, Zhejiang, the past decade has been his magic time. With the sunrise in sight and the world shrouded in an incredible blue tone, Unitree robots have gone from the laboratory demos to the Spring Festival Gala and then to the capital market.
However, for the embodied intelligence industry, the past decade has not been peaceful. Until the wave of AGI formed, the prospects of the global robotics industry were full of uncertainties. During this period, Boston Dynamics was even sold twice.
It can be said that Unitree has been supported all the way until its technical route was determined and its business model matured. It emerged from the horizon and coincided with the intelligent era. By looking through the prospectus, it can be found that in the past decade, about 40 venture capital institutions and industrial capitals have successively invested in Unitree. Meituan, Tencent, Alibaba, Sequoia China, Shunwei Capital, Matrix Partners, almost all the top institutions you can think of are on this list.
The vast majority of these institutions came in after Unitree's robotic dog appeared on the Spring Festival Gala in February 2021. This made the financing list of the company's subsequent rounds extremely long, indicating that everyone wants to get a ticket to the embodied intelligence field.
This makes us more interested in Wang Xingxing's angel investor. Borrowing from the recently popular TV series "The Leading Role", who was the one who discovered this entrepreneurial teenager with "two lights having wicks" before the curtain rose?
The story of a star entrepreneur is indeed attractive, but it is more meaningful to explore the environment that the entrepreneur has gone through: it more comprehensively explains the inevitability in the contingency of the entrepreneur's success and also presents the environment for Chinese entrepreneurs more comprehensively.
After in-depth analysis of Unitree Technology's prospectus, we found that in 2018, Jisi Investment became the angel investor of Unitree Technology. At this time, there were still three years until Unitree first appeared on the Spring Festival Gala.
01
A Master Without a School and a One - Second Investor
In the investment circle, Jisi Investment is not well - known. Tianyancha shows that it is an early - stage fund established by Variable Capital under GeekPark.
From a report in 2025, we can roughly trace the context of how Jisi Investment discovered Unitree.
Going back to 2017, Zhang Peng, the founder of GeekPark and the founding partner of Variable Capital, saw the prototype of Wang Xingxing's robotic dog, Laikago, in an obscure official account article. There was only a short video and a few pictures in it, and the content was very simple.
But Zhang Peng was very surprised.
This surprise had a reason. In 2015, Zhang Peng took domestic entrepreneurs to visit MIT. He saw a "robotic cheetah" made by a Korean professor in the laboratory. It was hung with steel cables and flying wires were in a mess, exuding a strong laboratory smell. Wang Xiaochuan in the same group took a photo with the cheetah's head. Everyone thought that this thing might become really powerful in another 10 or 20 years.
As a result, two years later, a Chinese team made something smaller, more mature, and closer to productization.
Zhang Peng flew to Hangzhou within three or four days. Wang Xingxing's company was in Binjiang District, Hangzhou. It was not a standard office building, but a somewhat shabby small building without a doorplate, and there was a factory building next to it. The office area was about 20 square meters, filled with parts and unassembled robotic dog components. You couldn't see any workstations when you went in. It was like a garage. There were two assembled robotic dogs. After they demonstrated the standard "come here and kick, and can't be kicked down", Zhang Peng found that there was no place to sit in the room. Finally, the two of them sat on the sofa in the corridor and talked for two or three hours.
Zhang Peng later described this meeting: "You could feel that he was not an entrepreneurship enthusiast who first set up all the entrepreneurship - related things and designed a sign and a logo. His company was just like a garage." What he was curious about was that this person didn't have a glorious academic background, didn't have partners from MIT or Caltech, and didn't have overseas professors as advisors. "He was not from an established school in the industry, but like a new master who suddenly appeared in the industry without a school."
Zhang Peng decided to promote this investment, and there was a technical judgment behind it: Boston Dynamics was taking the hydraulic route at that time, which was large in size and noisy. Although it looked cool, Zhang Peng thought this route was wrong - there were insurmountable problems with hydraulics, and electric drive was the trend. This was exactly the same as Wang Xingxing's judgment. In 2017 and 2018, intelligent electric vehicles began to emerge, Tesla was widely known, and full electrification was becoming a major trend. The technological dividend would surely spill over to more fields.
And globally, Wang Xingxing was the one with the highest degree of completion. He didn't come from a famous school and didn't have experience in successful projects, but his talent and focus on technology were obvious.
However, in 2018, when it came to actually paying the money, things were not that simple. At that time, the global robotics industry was in a slump. After being acquired by Google, the once - glorious Boston Dynamics suffered continuous losses, and Google sold it to SoftBank. Although the entire robotics track was cool, the commercialization path was vague, and it was almost impossible to see the possibility of making a profit. Mainstream capital withdrew collectively.
According to the plan, Unitree needed at least 15 to 20 million yuan in this round. Jisi Investment was an early - stage fund and couldn't provide that much. The first thing Zhang Peng did was to help Wang Xingxing find a leading investor. It took a lot of time, and they were once very close to reaching an agreement. There was even an investment letter of intent, but in the end, due to some changes, the other party didn't complete the investment.
Wang Xingxing was out of money. Later, Zhang Peng learned that Wang Xingxing took out his own savings to pay the employees' salaries without saying a word. He just implicitly said that there was almost no money in the account, which would affect the subsequent product delivery.
According to the framework, Jisi Investment was a follow - on investor and would give the money after the leading investor came in. But Zhang Peng knew that there was really no time to hesitate. "If we hesitated for another month, his company might go bankrupt." He was possessed by the "one - second partner" and decided to be the first one to transfer the money.
Soon, without signing any agreement, 2 million yuan from Jisi Investment was transferred to Unitree's account. That was the life - saving money for Unitree.
02
A Group of Bigwigs and a Small Fund
Jisi Investment is an early - stage fund with only 25 million yuan, and its scale is very small.
However, the LP team of this fund is extremely luxurious, including Chen Danian, Zhou Hang, Cao Yi, Liu Limeng, Zhao Xuejun, Jiang Xiaoping, Liang Jian, He Xiaopeng, and Zhang Tao, the founder of Dianping.
From a financial return perspective, for these bigwigs in the Chinese Internet industry, participating in such a small fund doesn't make much sense. Because the scale is really too small. Even Chen Danian, who contributed the most, only invested 5 million yuan.
Moreover, the closed - end period of early - stage funds is very long. There are not only successful cases like Unitree but also cases that fail halfway. By the time of actual exit, the return for each bigwig won't be much, probably just like pocket money.
But for them, "creating Unitree" is the meaning itself.
Before becoming bigwigs, these people were all entrepreneurs, or rather, innovators. After founding UC, He Xiaopeng became the founder of XPeng Motors. Zhou Hang founded Yidao Yongche, the earliest shared - mobility platform in China. Chen Danian is also a serial entrepreneur with a technical background. He founded Shanda Network and LinkSure Network and pioneered the online game and online literature industries.
Because for these people, innovation is more like a belief. Having witnessed it themselves, they also want to witness others' success.
Globally, it is this belief that enables the inter - generational transfer of innovation resources between China and the United States, making them the two poles of global innovation. In Silicon Valley, the culture of entrepreneurs helping each other is very strong. The PayPal Mafia sent a group of entrepreneurs to the power center of the technology industry; Paul Graham, the founder of YC, incubated Airbnb and Dropbox through early - stage investment. As generations of excellent entrepreneurs emerge in China, a self - inheritance system is also being formed.
The value of this inheritance system lies in lowering the threshold of "being seen". Before this, a young person without a famous - school background or work experience in a large company could hardly enter the vision of mainstream investment circles and industrialists. After being invested by Jisi Investment, Wang Xingxing found his "peers" there.
Wang Xingxing graduated from Zhejiang Sci - Tech University with a bachelor's degree. When taking the postgraduate entrance examination, he failed to get into Zhejiang University because of his poor English and was transferred to Shanghai University. He had few resources during his postgraduate studies. He designed the hardware and control algorithms from scratch and made the robotic dog XDog with industrial motors. He won the second prize in the Shanghai Robot Design Competition. At that time, it was still three years before MIT open - sourced the robotic dog algorithm.
He has an almost paranoid focus on engineering. He said that "in the three years of high school, his English only passed three times in all the exams", but this doesn't prevent him from having keen insight and implementation ability in engineering. When he made XDog, the most valuable motor driver was designed by himself. Later, when Unitree was established, the entire motor in the first robot produced was self - developed by Unitree. Then, the lithium - battery pack, wireless communication, vision module, and 3D lidar were also gradually self - developed and self - produced.
These technologies and products that outperform Boston Dynamics in the international market were successively born in an office like a garage. Over the years, even after several rounds of financing, Unitree has never moved.
In 2024, my colleague also went there, and it was exactly the same as what Zhang Peng described. It is said that He Xiaopeng also went to Unitree's office. He might have recalled his own entrepreneurial period there, which was also in a similar environment, or even worse, without an air - conditioner, where the prototype of XPeng was born.
So, where to start doesn't matter. As Zhou Hang summarized in "Re - understanding Entrepreneurship", "Acceleration is more important than initial velocity."
So, how can a startup get acceleration? Of course, someone has to give it a push.
03
The Power of "Cloud Computing"
After the 2 - million - yuan investment in 2018, Jisi Investment was deeply involved in Unitree's growth. Zhang Peng said in an interview in 2025: "In the past, the capital environment was good, and you had nine lives. It was okay even if you wasted eight and only had one left; now, you may only have half a life at the beginning. Entrepreneurship is more difficult, and the 'cloud computing' of the predecessors is even more important."
The so - called "cloud computing" is not computing power in the technical sense, but the transfer of resources, experience, connections, and judgment of the previous generation of entrepreneurs to the next generation. It cannot be completed by a single investment or a single meeting but is a long - term accompanying process over many years.
During the World Internet Conference in Wuzhen in 2017, GeekPark organized a gathering for entrepreneurs. Zhang Peng specifically asked Wang Xingxing to bring the robotic dog to introduce it to the Internet predecessors present. As a result, the robotic dog got stuck and crashed when crossing the threshold, and the scene was once embarrassing. However, the people present saw this thing for the first time and thought that this young man was quite amazing in doing this. They all encouraged him.
Although the demonstration failed, the seeds were sown. A few years later, many of the people present became investors in Unitree.
Not only Zhang Peng, but the LPs of Jisi Investment also have a deep connection with Unitree and are the earliest "computing power center" of Unitree.
Jisi Investment pioneered the "order - making system". The entrepreneurs invested by Jisi can name an LP to teach them entrepreneurship and business experience, and the GP will make a connection with this bigwig, organize a gathering, and facilitate the exchange. Practical bigwigs don't talk about strategies in a vague way but truly share the pitfalls and obstacles they encountered in the past. This kind of real experience is what entrepreneurs need the most.
In February 2025, a closed - door exchange meeting was held at the Huawei headquarters in Shenzhen. The on - site photos were full of the meaning of mentoring in the Chinese technology circle. There were Ren Zhengfei from the 1940s, He Xiaopeng from the 1970s, Yuan Jinhui, the CEO of Silicon - based Flow from the 1980s, and Wang Xingxing from the 1990s.
It is said that He Xiaopeng specifically brought Wang Xingxing. One generation passes on resources, experience, and connections to the next. This transfer is not written in any agreement but is the most fundamental operating system of the entire innovation ecosystem.
GeekPark calls this model "fixed - investment" - continuously discover, support, accompany, and then wait for time to give an answer.
The answer given by time is more abundant than you think.
Unitree Technology's operating income was 159 million yuan in 2023, increased to 392 million yuan in 2024, and soared to 1.708 billion yuan in 2025, a year - on - year increase of 335.36%. The net profit after deducting non - recurring gains and losses reached 600 million yuan, a year - on - year increase of 674.29%. In 2025, the shipment of its humanoid robots exceeded 5,500 units, and the total sales of quadruped robots exceeded 30,000 units. The gross profit margin climbed from 44.18% in 2022 to 60.27% in 2025.
This set of numbers together shows one thing: Unitree not only survived but is also becoming the most "calculating" company in the embodied intelligence track.
Meanwhile, Unitree's capital map is still expanding. The prospectus shows that Meituan holds a total of 9.65% of the shares through multiple entities and is the external shareholder with the highest shareholding ratio; Sequoia China holds 7.12% of the shares, and Matrix Partners holds 5.45% of the shares. After completing a nearly 1 - billion - yuan Series B2 financing in February 2024, giants such as China Mobile, Tencent, and Alibaba were introduced in the Series C financing. The valuation soared from 8 billion yuan to 13 billion yuan within half a year. As of before the listing, about 40 investors have successively entered the game, including national and local state - owned capitals, forming a complete map covering early - stage angels, top VCs, industrial capitals, and state - owned capitals.
A good company can speak for itself and will also attract more ears to listen. Wang Xingxing firmly believes that "when what you do conforms to the main theme of the era, some good resources will tilt towards you."
On the same day when Unitree passed the review, Jensen Huang, the CEO of NVIDIA, announced that the NVIDIA Isaac GR00T reference humanoid robot equipped with Sharpa Wave would be officially unveiled to developers. The body of this robot was created by Unitree.
This is a company born in a "garage" that has come into the spotlight of the world's artificial intelligence center. And in this journey, there is a group of people who have lit the street lights for it all the way.