Hot Money in Embodied Intelligence Shifts: Unitree's Stock Price Rebounds as Yinghe and a Wave of Robotics Companies Head Toward Bankruptcy Liquidation

Confidence in embodied intelligence seems to have returned.

Over the past three trading days, Unitree Technology's stock price has surged nearly 30% cumulatively, with its market value returning above the 200 billion yuan mark. Although it has not reached the high of over 400 billion yuan on its first day of listing, it is still a rare strong rebound in the past month, with a single-day gain once second only to its listing day.

Facing the continued rise in its stock price, Unitree Technology issued an announcement last night stating that the company's operations are normal and that there are no major matters that should be disclosed but have not been disclosed.

Unitree Technology announcement

Unitree also noted that the company's rolling price-to-earnings ratio has reached 388.25 times, close to 10 times the average level of the general equipment manufacturing industry. Meanwhile, the humanoid robot industry is still in the early stages of commercialization, and if implementation falls short of expectations, the company will face the risk of valuation regression.

As Unitree rebounds, another star robot company that was once equally high-profile is now facing bankruptcy liquidation.

On October 19, Yinghe Robotics held its first creditors' meeting after bankruptcy liquidation. Yinghe Robotics had a high starting point, having spun off from He Jianfeng's Yingfeng Group, and had successively received investment from ByteDance, GSR Ventures, and Meituan-affiliated enterprises, with a peak valuation of over $500 million.

But even so, Yinghe still could not escape the fate of bankruptcy, falling from its peak in just two years.

Yinghe is not an isolated case. Over the past two years, a large number of robot companies, including Zhicheng Dynamics, DeepBlue Robotics, CloudMinds Robotics, Longhui Medical, and Zhuineng Robotics, have gone bankrupt one after another, among which there is no shortage of star projects backed by industrial leaders, well-known investment institutions, and industry heavyweights.

On one side is a stock price rebound, while on the other is a quiet exit. The robotics industry has never lacked capital attention; it is just that the target of hot money has already changed to a new round. Even the difficulties faced by these robot companies are similar—the same beautiful vision, the same difficult implementation, and the same slowly dissipating confidence.

Beyond Unitree, more robot companies are still lining up and waiting anxiously for IPOs, and the time left for them seems to be running out.

The first batch of robot companies can no longer hold on

Contrary to people's conventional impressions, most of the robot companies that have gone bankrupt over the past two years are not nobodies; each has credentials.

Needless to say, Yinghe Robotics was backed by He Jianfeng, son of Midea's He Xiangjian, and received support from Yingfeng Group from the start. Founder Shen Gang is a heavyweight in the robotics industry and worked for many years at the Japanese robotics giant Fanuc.

Zhicheng Dynamics, which focuses on pool robots, can be seen as a representative of star industrial projects. Founder Pu Hanlai came from Microsoft Asia Engineering Institute and once led the Xiaomi ecosystem company Moxiang Technology; the core team all came from companies such as Microsoft, Xiaomi, and DJI, and last year it also received a strategic investment of tens of millions of yuan from Ecovacs.

CloudMinds Robotics, founded in 2015, even secured multiple rounds of support from SoftBank China and Foxconn, and at one point planned to list in the United States. DeepBlue Robotics, backed by DeepBlue Technology, also previously disclosed that it had won an order worth over 700 million yuan from South Korea's LG and planned to deliver tens of thousands of robots within three years.

But none of this could support them in continuing. Like many highly anticipated embodied intelligence projects today, most of them obtained huge financing and expected orders in the early stages, but after several years they were still unable to deliver and never secured a stable source of revenue, surviving only on financing.

In 2021, Yinghe won an order worth nearly 100 million yuan from Foshan Lighting to build a smart manufacturing workshop project for bulb lamps. But the project was still not accepted two years after the deadline. The two sides eventually went to court; Yinghe sought about 104 million yuan but failed, and was instead ordered to return 26.904 million yuan and dismantle the equipment at its own expense.

The LG order that DeepBlue Robotics heavily promoted also failed to be executed on a large scale. According to previous media reports, after demand in the disinfection robot market receded, the order was not executed. Relevant employees told the media that "many products cannot be effectively implemented in real scenarios and can only be used for display."

Products were never able to be implemented, yet both companies expanded aggressively in their early stages, with bloated business lines. Just one year after its founding, Yinghe secured a research and development production base in Shunde with an investment scale of 10 billion yuan and an area of 500 mu, claiming it would use "robots to build robots."

DeepBlue, meanwhile, ventured into robot products in multiple fields including cleaning, disinfection, delivery, lawn mowing, and even passenger-carrying AGVs. At the same time that DeepBlue's Changzhou factory owed wages, DeepBlue was still putting its second factory into production in Shangqiu, with planned capacity of 20,000 units.

CloudMinds founder Huang Xiaoqing later reflected in an interview that the biggest mistake he made in starting the business was not enabling the company to achieve cash-generating capability as soon as possible. "To put it bluntly, you have to do things that make money and first let the company replenish its own blood." According to media reports, at its peak CloudMinds had more than 800 employees and planned as many as six product lines, consuming a lot of funds.

While being unable to achieve self-sustaining cash flow for a long time, investors' interest in these early robot projects was also gradually weakening. Unlike the currently popular embodied intelligence concept, most early robot projects focused on segmented service scenarios such as cleaning and delivery, and were inferior to general-purpose robots in both market imagination and technological narrative.

Huang Xiaoqing said that starting in 2024, their financing became especially difficult. One reason was that many startups had newly flooded into the humanoid robot track, and investors could invest in early-stage projects at lower valuations, while late-stage high-valuation projects like CloudMinds were instead ignored. At that time, CloudMinds had already fallen into an operational crisis due to interrupted financing and began shrinking its team and offices.

Not only domestically—overseas star robot companies have also failed to escape funding difficulties. In July this year, U.S. surgical robot company Vicarious Surgical initiated bankruptcy liquidation. Founded in 2014, the company had won support from well-known figures such as Bill Gates and Jerry Yang, and was valued at $1.1 billion when it went public via a reverse merger in 2021. But until this year, its product had still not been approved for commercialization, and it ultimately had to bow out of the market.

Hot money surges toward newcomers

Most of the previous wave of robot companies waited in vain for financing, but hot money is still flowing into the robot track.

IT Juzi statistics show that in the first half of 2026, total domestic embodied intelligence financing exceeded 93 billion RMB, a fivefold surge compared with the same period last year. As many as 22 embodied intelligence companies are valued at over 10 billion, compared with only 3 last year.

IT Juzi data

The financing speed of new companies has already left their predecessors far behind. CloudMinds spent nearly a decade and went through five rounds of financing, raising about 4 billion yuan in total. Zhi Pingfang raised nearly 5 billion yuan in a series of financings in the first half of this year, while Zibianliang completed four consecutive financing rounds within two months. Both claim to be the first embodied intelligence company in the Greater Bay Area with a valuation of 20 billion yuan.

A bubble seems inevitable, but no one is willing to stop. According to incomplete statistics from Yicai, in the first half of this year, Sequoia China participated in at least 20 embodied intelligence financing deals, investing in 14 companies. Hillhouse Venture Capital participated in at least 18 financing deals, also investing in 14 companies. Sometimes they even appeared in clusters, co-leading investments.

The moment of hottest market sentiment was undoubtedly Unitree Technology's first day of listing. That day, Unitree Technology's share price rose as much as 629.44% intraday, and its market value at one point approached 450 billion yuan—higher than the combined market value of the four automakers NIO, Li Auto, XPeng, and Leapmotor—while Unitree Technology's revenue last year was only 1.7 billion yuan.

After Unitree's success, investors may have become even more frenzied. According to incomplete statistics, as many as more than 20 embodied intelligence companies are aiming for an IPO this year alone, including not only complete-machine players such as Zhiyuan Robotics and Galaxy General, but also a large number of players in segmented fields such as embodied models and dexterous hands.

But this listing boom quickly cooled.

The Information previously reported that the China Securities Regulatory Commission has given informal guidance to some investment banks and investment institutions to raise the approval threshold for IPOs of humanoid robot companies. Reuters reported that regulators are particularly concerned about companies' inflated valuations and whether revenue reliant on government projects truly reflects commercial demand.

Yunshenchu, whose IPO was accepted in May this year, had revenue of 340 million yuan last year and net profit attributable to the parent company of 28.68 million yuan, turning a loss into a profit. In the first half of this year, Yunshenchu's revenue doubled year on year, but it returned to a loss.

The prospectus shows that in the past two years, Yunshenchu sold only 4 units of its DR series humanoid robots in total. In the Shanghai Stock Exchange's first round of inquiries, particular attention was paid to issues such as whether Yunshenchu's revenue is sustainable, when humanoid robots will be sold at scale, and the structure of orders.

Some star companies have begun to postpone their listing plans. As early as July this year, Zhiyuan announced that it had started preparations for a Hong Kong stock listing. At that time, investors expected the company could be listed as soon as August. But several months later, Zhiyuan had still not listed smoothly as expected.

Caijing reported last week that Zhiyuan instead sought a new round of financing of about 1 billion to 1.5 billion yuan at a valuation of about 50 billion yuan, and that the round was nearing completion.

According to Hong Kong Stock Exchange requirements, if a listing applicant completes new financing after submitting its application but before listing, unless a special exemption applies, it must wait 120 days after the financing is completed before it can be listed and traded. In other words, Zhiyuan has no hope of an IPO within the year.

HKEX "Guide for New Listing Applicants"

But even as IPOs cool, investors remain enthusiastic. As Unitree Technology's share price rebounded, many embodied intelligence companies have recently completed multiple financing rounds. Stardust Intelligence announced on October 13 the completion of a new round of financing worth 1 billion yuan, with the company valued at 15 billion yuan. Fa'ao Robotics also announced in October that it had completed consecutive C+ and D rounds of financing, totaling nearly 1 billion yuan.

Who is paying for robots?

Even though most of the new wave of embodied intelligence companies have grander visions and more compelling demonstrations than the previous generation of robot companies, the problems they face are similar to some extent: how to truly achieve commercialization and how to continuously generate their own cash flow.

Counterpoint statistics show that in the first half of this year, global humanoid robot shipments exceeded 22,000 units, up nearly 300% year on year. But entertainment and commercial performances, scientific research, and data collection still accounted for more than 60%, while intelligent manufacturing and warehousing logistics accounted for only 13% and 5%, respectively. In other words, robots truly used in real production remain a minority, and most robots still provide emotional value.

Just like the domestic robots that appeared en masse on this year's Spring Festival Gala, they are more keen on performing than producing. Even these performances are costly: industry lore has it that each robot company appearing on the Spring Festival Gala pays a cooperation fee of 100 million yuan—roughly equivalent to more than half of Unitree's R&D investment last year.

Previous media reports said that a more important source of revenue for some robot companies is data collection factories in cooperation with local governments. Data collection is essentially a kind of resource reuse within embodied intelligence: robot companies sell robots to data collection factories, then spend money to buy back data from the data collection centers.

In September this year, MegaRobo founder Shao Tianlan posted on WeChat Moments, attacking the domestic embodied intelligence sector for having a batch of "assembled-deal" companies that heavily use related-party transactions with "data collection centers" to manufacture fake, unsustainable revenue.

Shao Tianlan's WeChat Moments post

In his WeChat Moments post, Shao Tianlan named Galbot. Galbot responded publicly that it does not engage in pointless verbal sparring and will not be disturbed by short-term noise.

Data collection is indeed one of the main sources of revenue for some embodied intelligence companies. Leju Robotics' previously disclosed prospectus shows that in 2025 the company sold 577 units of its full-size humanoid robot Kuafu series, generating revenue of 178 million yuan, of which about 45% came from the data collection sector, while revenue from industrial manufacturing accounted for 3.88%.

By contrast, progress in getting robots into factories is much slower. Even after entering production lines, the work robots perform is quite limited.

Overseas robotics company Figure disclosed that its robots have cumulatively operated for more than 1,250 hours at BMW's U.S. plant, moving over 90,000 parts and participating in the production of more than 30,000 vehicles.

That sounds impressive, but what it specifically does is take sheet metal parts off racks, place them onto welding fixtures, and then hand them over to traditional industrial robots to complete the welding. Around this single action, the two sides spent months testing and debugging.

The International Federation of Robotics noted in a report released in September this year that humanoid robot applications are still mainly specialized tasks and often require manual teleoperation; training and maintenance costs are too high, and the economic returns of industrial applications are insufficient.

Apart from some leading companies, more robot companies may not even have real delivered orders. Counterpoint statistics show that in the first half of this year, the top five manufacturers in the global humanoid robot market together accounted for 86% of total shipments, with the top two companies, AgiBot and Unitree, holding 43% and 31% market share respectively.

Counterpoint statistics

Capital is also concentrating toward the leaders. IT Juzi statistics show that in the first half of this year, the top 20 companies by financing amount in the domestic embodied intelligence sector together raised 55 billion yuan, accounting for 59% of the industry's total financing.

Even companies that have already secured more orders have not escaped losses. UBTech sold 921 full-size humanoid robots in the first half of this year, with related revenue reaching 590 million yuan, making it the company's largest business. Over the same period, the company's total revenue doubled to 1.269 billion yuan, but it still lost 339 million yuan.

Amid the current surging investment boom in embodied intelligence, the new batch of robot companies does not yet need to worry about survival, just like those last-generation robot projects that were once riding high a few years ago.

But market confidence may not last too long either.

This article is from the WeChat public account "Shanshang," author: Xue Xingxing, editor: He Jian, published by 36Kr with authorization.