Dadao Technology's One-Year Anniversary of Decline: The Death of an Embodied Unicorn

If you’re trading stocks, rely on Jinqilin analysts’ research reports—authoritative, professional, timely, and comprehensive—to help you uncover high-potential theme opportunities!

（Source: Leifeng.com）

“Embodied bubbles spawned by capital of the Chinese-style.”

Author丨Qi Chengyong

Editor丨Lin Juedmin

In the spring of 2026, the embodied AI industry is迎接ing a wave of capitalization. Unitree Technology is rushing to list on the A-share market, Zhyun Robot is preparing for a Hong Kong listing, Xinghai Tu, Fourier Intelligence, and others have completed shareholding restructuring, and many embodied companies’ valuations have already surpassed 100k yuan. For these companies, getting investors on board seems to have become the go-to rule for increasing their sense of security.

But from a historical perspective, high valuations do not necessarily mean stability. The most typical case is the once-highly valued embodied unicorn—Daduan Robotics.

This embodied “castle” was built by founder Huang Xiaoqing over 10 years. Its valuation once exceeded 20 billion yuan, yet more than a year ago it experienced a sudden collapse. So, who “killed” this embodied unicorn?

First, let’s talk about a detail that goes against conventional thinking. On social media, there’s a widely circulated version: Daduan Robotics died from a broken capital chain, employee lawsuits, and debt disputes. But in fact, as of March 2025, the total amount Daduan was subject to enforcement under judicial records was only 35.3 million yuan. For a company founded in 2015, through seven rounds of financing, whose valuation once reached 22.3 billion yuan—the “leading unicorn in China’s robotics industry”—this is not a huge number. You have to know that this company’s real estate assets—including the office building of the Guangzhou branch and hundreds of acres of industrial park land—are all self-owned assets, and selling them could help Daduan get out of trouble. So why did it still end up in a failed, stop-start situation where the team scattered and the founder left to go to Hong Kong to found a new company?

Therefore, the key trigger for the collapse of Daduan’s embodied “mansion” does not come from downstream, but from tracing back to the founder Huang Xiaoqing’s equity structure and Daduan’s upstream structure.

01

Limitations of the first generation of embodied founders

If we divide China’s embodied field entrepreneurs by generations, we can roughly categorize them into three generations.

The third generation is represented by engineers or scientists born in the 1980s, such as Unitree Wang Xingxing, Zhyun’s Li Zhijun, Xingdong Yuan’s Chen Jianyu, and others. In their eyes, algorithms and models are the moat. Robots are not fixed control logic; rather, motion strategies are learned. In comparison, hardware is not the most important.

The second generation is represented by people born in the 1970s, such as UBTECH Zhou Jian and Agile Robot Li Tong. Compared with the third generation, they place more trust in “traditional control + limited intelligence,” treating the robot body as a “product,” while the business model is “selling equipment.”

As for the representatives of the first generation of robot entrepreneurs, most are born in the 1960s. To a certain extent, they carry heavy limitations of the era. On the one hand, they often believe robots are merely executors and underestimate the complexity of physical interaction. On the other hand, they rely too much on “relationship-driven” approaches—for example, government orders and state-owned financing. Huang Xiaoqing is a typical representative of this generation.

Before founding Daduan, Huang Xiaoqing was a technical executive with an excellent resume.

After graduation, he spent years deeply in the electronic communications field. He worked for five years at Bell Labs. In a place revered as a hall of electronic communications science, Huang Xiaoqing led the development of a brand-new system that compressed seven cumbersome subsystems into three—earning fame in a single battle and winning a good reputation inside the lab.

In 1995, he returned to China briefly, bringing in a group of Bell Labs colleagues to join UT Starcom. Then he went back and forth between Silicon Valley and Hangzhou. At that time, China’s electronic communications industry was slowly moving into the 3G era. With a sharp technical sense, Huang Xiaoqing quickly founded a new company, Wacos, aimed at empowering the 3G era. But because the concept was too advanced, the “Wacos 3G” project ultimately failed.

Until 2007, China Mobile extended an olive branch to Huang Xiaoqing and offered him the position of president of the China Mobile Research Institute. After spending 8 years at China Mobile, Huang Xiaoqing witnessed China’s evolution from 3G to 4G firsthand.

In 2015, as the 5G era began to take shape, Huang Xiaoqing sensed opportunities and transformation again. He had the idea of starting a business, but wasn’t sure exactly what to do. Before founding Daduan, Huang Xiaoqing had been in the communications industry for a long time and had no professional experience in the robotics field.

In Huang Xiaoqing’s own account, the reason he chose to start a robot company came from his science-fiction dream. Huang Xiaoqing really liked Star Trek. In the show, there’s a robot named “Data.” Inspired by this, he combined cloud communications with robots and proposed the concept of “cloud robots.”

However, more former Daduan employees believe Huang Xiaoqing was not that innovative. He was only able to pick the track because he often worked across oceans, had a keen outlook, came into contact early with Professor James Kuffner’s “cloud robotics” concept at Carnegie Mellon University, brought it into China, and belonged to a “localization team.”

02

The “communication gene” of robots

With a communications background, plus the popularity of the concept of “cloud intelligence” at the time, Daduan Technology quickly secured funding.

In 2016, Daduan completed its first round of financing, with SoftBank Group and Foxconn among the investors, jointly investing $30 million. At that time, the person steering Foxconn—its parent company Hon Hai—was Guo Taiming. The following year, SoftBank Group increased its commitment again, jointly investing $100 million with ZGC Development Group. This transaction set a record in the cloud intelligence sector at the time.

Daduan started off very smoothly and, within a year, grew into an industry unicorn.

But in reality, so much financing did not lead to major breakthroughs in product development and technology. In its planning, Daduan set multiple robot scenarios, including the XR-series general-purpose robots (XR generation 1 to 4), the Ginger Lite robots for low-cost front-desk reception and service, and unmanned retail robots, among others.

Huang Xiaoqing once envisioned that the shipment volume of unmanned retail robots could reach the tens of millions, even tens of millions of units. For reception and service robots, he expected at least 100k units shipped per year. By 2026, however, Galaxy General, which is known for unmanned retail robots, had not met the rollout data Huang Xiaoqing expected at the time.

More awkwardly, not long after Daduan received investment, it began to play with capital tactics—trying to boost revenue and go public in the U.S. stock market.

According to recollections from early Daduan employees, Daduan “didn’t get anything done early on, but instead made a lot of phones, including encrypted phones specifically for enterprises. They were originally supposed to be sold to Foxconn, but it didn’t work out.”

In 2017–2018, Huang Xiaoqing faced difficulties as the robotics business encountered setbacks, while he was also considering listing on Nasdaq. On one hand, he needed orders to account to investors; on the other, he needed performance data to support the valuation. That’s why he was forced to make phones. After all, the communications field was his main base.

This history was deliberately downplayed, but it left “genes” in Daduan: a communications way of thinking. Huang Xiaoqing forcibly transplanted China Mobile’s “operator” mindset into the robotics industry, fantasizing about becoming “China Mobile in robotics”—not selling robots, but selling services of “cloud brains,” charging by traffic.

03

The capital player Huang Xiaoqing

In 2019, Daduan tried to list in the U.S. At first, it planned to go through Nasdaq, but it ran into the NYSE抢单 situation—Huang Xiaoqing’s fees were immediately waived by several tens of millions of yuan. So Huang Xiaoqing, without hesitation, turned to the NYSE and planned to raise $500 million. After the prospectus was issued, investors didn’t believe it, but the U.S. government did—so it ended up on the sanctions list.

This IPO was aborted due to the “Entity List” sanctions, but the U.S. government’s backing ironically turned it into “a blessing in disguise.” After being forced back to China, Daduan began its “招商” journey.

Soon after Daduan’s failed NYSE listing, the Shenzhen government and the Shanghai government stepped in one after another. Daduan then moved to Shanghai and was also granted “more than a dozen billion” along with 243 mu of land and many preferential terms.

During this process, several key figures in Daduan’s management team played important roles, including Yang Guanghua and Wang Bing. They mainly handled To G business at Daduan: Yang Guanghua had worked with Huang Xiaoqing since the UT Starcom days; Wang Bing had previously served as vice president of Beijing Kingdee Software. Some peers described them as “good at dealing with government-business relationships.”

This model driven by government-business ties helped Daduan’s capital lineup reach its peak in 2021–2023, but its equity was also highly fragmented: Shanghai Chengtou, Zhuhai state assets, Guangzhou Zhishicheng, Ganzhou Nankang District Chengfa Group… Among the 67 shareholders Daduan has penetrated, enterprise shareholders make up 57.

Such an equity structure is not unique to Daduan. In 2024, embodied intelligence companies that completed large-scale financing also have shareholders including Tencent, BYD, Hillhouse, Sequoia, and local government funds. The list is longer, but the essence is the same—“a mixed platter”—and they all hope to build a shareholder structure that can support ongoing financing.

In China’s LP structure, government-guided funds occupy an important place. Daduan’s predicament was that “no next LP was willing to take over the position.” The luck of today’s leading companies is that “there is still a next LP observing.” But the due diligence logic of the “next LP” is exactly the same as that of the “previous LP”: they want orders, revenue, and certainty.

Now it looks like the leading embodied companies that got financing early and had the Snowball rolling in 2024, by 2026, will face the same tests of orders, revenue, and certainty. Many embodied companies’ finance leaders’ top KPI this year is to secure mass production data and clarify revenue expectations, because they’ve also reached the stage where they need to prove commercialization.

04

Glory

In 2021, Daduan arrived at its most glorious period.

At the time, there were 12 PMOs (project management offices), and the number of projects reached a peak. Huang Xiaoqing held regular project reporting meetings. Humanoid robots, cloud brains, smart joints, digital humans, delivery robots, agricultural robots, security robots… Each project lead had PPTs, had data, and had promises of “nearing mass production.”

Back then, Daduan even had its own factory producing motors—not only controlling costs, but also enabling customization. In the 2021 timeline, Daduan’s momentum surpassed Unitree, which had not yet risen at the time.

“Everyone had tons of project products, and the sales formats we could support were also especially diverse.” One former PMO member recalled. He later went to another embodied intelligence company and found that its PMO also had seven or eight, and the founder also held regular project reporting meetings; project progress charts were also posted on the wall—the red represented delays, and the yellow represented risks. The proportions were almost the same as Daduan in 2021.

Daduan’s six product lines at the time were later regarded by embodied entrepreneurs as “lessons”—“don’t be greedy, focus.” But the real situation is that focusing is both a risk and a luxury.

Focusing means going all-in on a single track. Whether with the public or with investors, they will keep a close eye on that business line. The risk comes from the possibility that you could “pick the wrong bet.” In such an unclear and murky field as embodied robotics, blindly choosing any one route may cause you to fall behind the wind. In addition, if the business line isn’t broad enough, it also means the product lines are insufficient to support valuation.

Now, leading companies are quietly expanding their product lines as well—from humanoids to quadrupeds, from industry to home, from hardware to software, from robots to “embodied intelligence large models.” They don’t say they are expanding; they say they are “building an ecosystem”—which is no different from what Daduan called “the cloud robotics ecosystem” in 2019.

05

Huang Xiaoqing’s foresight

The founder’s vision is not behind.

In today’s embodied field, many investors discuss how the entrepreneur’s vision is how it is. But from this angle, Huang Xiaoqing is almost an entrepreneur with an extremely strong technical vision.

In 2020, Huang Xiaoqing mentioned in multiple talks his thoughts on humanoid robots: “From a psychological perspective, humans can’t accept a table having a conversation with us. Robots need to share humans’ space and share humans’ tools.”

At the time, nobody understood. In 2020, humanoid robots were still exclusively for science-fiction movies. Boston Dynamics’ Atlas was still flipping around in the lab, and Unitree’s quadruped robot had just sold its first unit.

Four years later, at the 2024 World Artificial Intelligence Conference, a founder of an embodied intelligence company expressed a similar view in a similar setting: “Humanoid is the ultimate form, because only a humanoid can use all the tools created by humans.” The applause from the audience thundered. This founder was nearly 30 years younger than Huang Xiaoqing, wearing the same black T-shirt, saying the same words.

The difference is: when Huang Xiaoqing said it in 2020, Daduan’s humanoid robots were still on paper; when this founder said it in 2024, his humanoid robots could walk and jump, although they couldn’t work yet.

In terms of ideas, Huang Xiaoqing was also ahead of his time.

Long before “cloud brains,” Huang Xiaoqing also realized the importance of reducing the cost of joints. In 2020, Huang Xiaoqing said in a media interview: “In the future, the number of robot joints determines its performance. Joint costs must come down, while maintaining high quality—only then can the robot era be started.” Not long after that, Daduan established a production and R&D base in Shanghai, hoping to bring joint costs below 1,000 yuan.

In 2025, many embodied companies only started to realize that bringing down the price of joints and dexterous hands is the way out—and Huang Xiaoqing had this idea 5 years earlier.

Huang Xiaoqing thought so far ahead that the investors in 2020 couldn’t understand it, the state-owned LPs in 2021 didn’t dare to invest, and the Nanjing industry fund in 2023 felt “the pace of technological commercialization does not meet the standard.” He fell because he was “too early,” but being “too early” itself isn’t wrong—if he could have lasted until 2024, he would have been a “prophet.”

But in a sense, are the current leading companies also “too early” now?

Huang Xiaoqing couldn’t achieve it partly because his mass production plan kept missing targets, and partly because his equity structure left Daduan no chance to survive until the day of execution.

06

The full story of the downfall

In the summer of 2023, at the WAIC conference, Daduan exhibited. In an environment where almost everyone was doing demo performances, the robot drew attention as it tried to play basketball in front of the whole crowd—but missed, and the scene was filled with awkward silence and disappointment. Daduan had hoped to prove its “big brother” status, but it ended up making a fool of itself.

After that, when Daduan returned to public attention again, there were almost no positive messages. In early 2024, Daduan began to have unpaid wages.

“In late January 2024, they were still preparing intensely for the annual meeting. By the 31st, suddenly they couldn’t pay salaries anymore.” A former Daduan employee recalled. At the time, the explanation was a financial system failure. On February 8, one day before Chinese New Eve, Daduan urgently convened an all-staff meeting on Feishu. For wages, the portion above 10k yuan was paid at half. At this meeting, Huang Xiaoqing did not attend. Only two people from management participated: one was the HR负责人 at the time, and the other was the CFO. After executing the plan for two months, salaries completely stopped in March and April, and the HC system was fully frozen. In May, after all employees received 10k yuan, the housing fund and social insurance were stopped one after another.

What was Huang Xiaoqing doing? According to his own account, during that year he met investment institutions “more than the total of the past nine years.” He went to Hefei, Tianjin, and Xiamen; farthest north, he even went to Jilin. For these local investment institutions, he wasn’t just meeting them once quickly—he held countless meetings. “No matter whether they invest or not, I did a very important educational session for them.”

But in fact, according to investors, the content of Huang Xiaoqing’s “education” was almost the same as when he went to the U.S. to list in 2019: cloud brains, smart joints, humanoid robots, and home service.

Five years later, the formatting of the PPT changed, but the core logic didn’t.

“In 2024, the only money people can truly raise in the market is government money managed by private GP managers.” This was his conclusion. The subtext of this sentence is: private GPs don’t have money, and government money doesn’t dare to invest, and Daduan happened to get stuck in that gap.

Daduan’s Series C round was co-led by Guangzhou Zhishicheng Group and Shanghai Guosheng Investment Group, raising over 1 billion yuan. Only half a year later, the capital chain broke. Huang Xiaoqing said that in October 2023, a certain institution was originally supposed to deliver 300 million yuan, but it failed to settle.

Huang Xiaoqing didn’t name the institution, but as long as you understand Daduan’s financing history, it’s not hard to tell: 300 million yuan is a typical investment size for a local government industry fund. In 2023, Daduan had planned to list in Hong Kong and raise $500 million. The approval was temporarily delayed due to “technology commercialization progress not meeting the standard”—and it was the Nanjing industry fund, a typical local government industry fund, that did it.

If you pay attention to today’s embodied financing field, you’ll find that the current top start-ups are taking money from Tencent, Meituan, and Jinshi Investment. They’re also looking for “money that can tolerate uncertainty.”

But in China, even industrial capital like Tencent and Meituan still can’t avoid the presence of government-guided funds in their LP structures. Therefore, this is not a predicament unique to Daduan.

07

Post-mortem of the downfall

To understand Daduan’s death, you have to go back to Huang Xiaoqing’s background—UT Starcom.

In the 1990s, as a co-founder and CTO, Huang Xiaoqing participated in founding UT Starcom. In the era of pagers (“Xiao Ling Tong”), this company was once a “dominant force along the Binfjiang River.” It had a market cap of 10 billion USD and annual revenue in the hundreds of billions, posing a threat even to Huawei. But UT’s decline at its core was internal politics.

According to recollections from former UT employees, internal “company politics” emerged in UT. Zhou Shaoning (former Google Greater China president) and Huang Xiaoqing had different ideas: Zhou Shaoning was good at making money, doing R&D, and making products. The two “would have power struggles.” After Zhou Shaoning left, Huang Xiaoqing took over, but “Huang Xiaoqing is relationship-driven; he understands some technology, but he isn’t particularly good at it.”

This “relationship-based” management style was replicated completely in Daduan as well.

An employee who joined Daduan in 2018 and managed the middle office said: “The chaos inside Daduan has always been like this. Mini (Huang Xiaoqing’s wife, who served as CFO) especially enjoys it when people go to report on others, saying who is not good. But if you try to say who is good, they may not listen. Also, they like to take credit. They often say, ‘Thanks to me xx’—basically, all the achievements are theirs, and all the problems are someone else’s.”

Huang Xiaoqing’s “one-man rule” style was even more extreme. “No matter how high a position you hold, no matter how many employees you have, in many things you really do need to go all the way down and figure out all the details yourself.” That’s how he described it in an interview, but what employees saw as fact was: “Huang Xiaoqing manages everything. Many small and trivial things need his attention. Working over there is really annoying.” Others also mentioned that when Huang Xiaoqing held meetings, he would directly scold people: “You people, these xx,” “These things are so simple, xxxx,” and his language was harsh.

Beyond detail management, there were also investors who commented on his overall style: “Bill and Mini (Huang Xiaoqing’s couple) are both pretty good people, but their biggest problem is lack of business sense. Both of them are stubbornly clinging to a technical route. Mini has improved a bit these past two years, but Bill is still like that.”

More destructive was the hiring mechanism. In Daduan’s later stage, it introduced many executives with backgrounds from Alibaba and ByteDance. “Not very practical; they love formalism, upward management, and all kinds of weekly reports and daily reports. They report twice a day.” But when problems occurred, they didn’t directly solve them—they reported up layer after layer. “Everyone knows what kind of boss he is. Everyone keeps making all kinds of reports to the boss, but the boss doesn’t have time to sit down and think about what the company’s next operating strategy is.”

The experience of Zheng Qing basically reflected the atmosphere at Daduan. This pillar from UBTECH Walker’s team joined Daduan in 2022. He found that “the technical route and architecture were all set by Huang Xiaoqing; nobody can change them.” He ultimately chose to leave, and then co-founded Yuanluo Technology with his high school classmate Lian WenZhao—later, the latter had worked at Figure AI and obtained a large round of financing in 2025.

According to one former Daduan employee’s summary, there are four kinds of people in the company: first, people with relatively high tolerance, but once doers are suppressed to a certain extent, it becomes hard for them to keep going; second, people with exceptional ability who follow the motto “Where I’m not tolerated, there’s somewhere else that will tolerate me”—“I’m not going to waste time with you anymore; I’ll leave”; third, “Well, you can say whatever you want, you can do whatever you want. But I also have my own principles”; fourth, “Anyway I don’t have any ideas. You do whatever makes you happy—I’m here just to serve you. As for the people below, I don’t need to manage those things.” In Daduan’s later period, most of the people left behind were the fourth kind.

08

Rebirth and lessons

In April 2025, a video showed Huang Xiaoqing, as the chairman of Hong Kong-style robots, accepting an interview from Hong Kong media.

He was in his 60s, with gray hair, but he still said the same things: “home service,” “cloud brains,” and “100k yuan robots.” The backdrop changed—from the 243-mu base in Shanghai to a small office in a Hong Kong office building; his identity changed—from founder of Daduan to chief scientist of Hong Kong-style robots—but the story didn’t change.

Hong Kong-style robots is a joint venture between Daduan and Guohua Group, a Hong Kong-listed company, with Guohua holding controlling interest. Guohua promised that in the next one or two years there would be financing, but the condition was that “the team must build the business, technology, and products”—and this requirement is exactly the same as what investors asked Huang Xiaoqing when Daduan was just established in 2015.

In addition, in today’s pitch-deck PPTs for large-scale financing in embodied companies, there are also words like “home service,” “cloud brains,” and “100k yuan robots.” His investors are the same group of people as those who invested in Huang Xiaoqing back then—just with different fund names.

The difference is: Huang Xiaoqing is in his 60s, while that founder is in his 30s.

Daduan’s death is not an endpoint, but a cycle. This industry is repeating the same story: technology ahead of time, grand scenarios, financing-driven growth, government connections, expansion of product lines, strong founder control, breakage of the capital chain, unpaid wages, layoffs, death, rebirth, and then another death.

Today’s leading companies face similar structural dilemmas as Daduan. Their luck is that Daduan died in 2024 while they lived until 2026. But what happens after 2026?

“Who killed the once embodied intelligence unicorn?”

First, it’s not technology and route—Daduan’s cloud brains, smart joints, and humanoid robots didn’t have an outdated technological route. Today’s leading companies are still taking the same path. Second, it’s not the market— the humanoid robot track saw a boom in 2024–2025, and real market demand exists. Today’s leading companies are still telling the same story. Third, it’s not management—Huang Xiaoqing’s “one-man rule” really is a problem, but now, the founders of leading companies are almost all “technical dictators,” with styles essentially identical to Huang Xiaoqing’s.

The real reason is that the logic of this industry hasn’t changed.

Technology ahead of time, grand scenarios, financing-driven growth, government connections—this logic made Daduan in 2015, dragged Daduan down in 2019, created a new unicorn in 2024, and will also test them at some point in the future.

Daduan is not a “failure.” It is a “pioneer.” The mistakes it made are being repeated by today’s companies. The path it took is the one today’s companies are taking as well. The warning Daduan gives to all embodied intelligence companies is that they must stay clear-headed and understand how to find the variable that will allow them to survive until “execution.”

What is that variable? Daduan didn’t find it. Maybe the leading companies of 2026 are looking for it too.