Incubated by the second generation of Midea, with ByteDance and Meituan taking equity stakes: How did Yinghe Robotics end up in bankruptcy?

On October 19, Yinhe Robotics will hold its first creditors' meeting.

As reported by Southern Metropolis Daily, the Shenzhen Intermediate People's Court accepted the bankruptcy liquidation application filed by creditors against Yinhe (Shenzhen) Robotics and Automation Technology Co., Ltd. in July this year. Founded in 2020, the company is now entering the stage of sorting out its assets, debts and repayment arrangements.

Five years ago, the discussions around it were about a completely different set of figures.

In January 2021, when Investment Post reported that Yinhe had received investments from ByteDance, Meituan-affiliated enterprises and other investors, an investment professional close to the project revealed that the company had been established for only ten months with a valuation of USD 500 million, and many institutions were competing to contact it. At that time, the investor also mentioned that the company's specific products had not been officially launched yet.

From investment institutions scrambling to enter the market to creditors waiting for repayment, Yinhe has gone through six years.

Its experience is worth inquiring precisely because this company once had conditions that many entrepreneurs could hardly obtain: a founder with rich industrial experience, a strong incubator, and well-known investment institutions that entered one after another. These conditions helped Yinhe quickly obtain funds and projects, but failed to allow it to smoothly go through the complete chain of delivery, acceptance and payment collection.

01 Yinhe's starting point was indeed not low

In March 2020, Yinhe announced its establishment. Reports from Nanfang Daily that year introduced it as a first-level subsidiary under the Infore Group, with Shen Gang as the founder and CEO. It initially targeted robot products in the fields of emergency management and urban governance, and carried out business in combination with intelligent manufacturing.

Infore Group was founded by He Jianfeng. As the son of He Xiangjian, the founder of Midea, He Jianfeng's family background made this project attract much attention from the very beginning.

Shen Gang brought another kind of appeal to the project. He went to Japan for further study in his early years, joined FANUC after obtaining his doctorate, and returned to China in 2014 to serve as a director and executive deputy general manager of Shanghai FANUC. Later, he joined Country Garden to participate in the preparation of Bozhilin Robotics, and entered the Infore system after leaving the position. He has direct hands-on experience in robot R&D, market expansion, and industrial projects of large enterprises.

For investment institutions, such a combination easily arouses expectations: there are people leading the work with technology and industrial experience, corporate resources and capital support are also guaranteed, and funds poured in later.

According to the financing record publicly disclosed by Li Lun, partner of Panda Capital, shareholder and director of Yinhe, in 2026, Yinhe's angel round financing in 2020 exceeded 90 million yuan, the Series A round in 2021 exceeded 400 million yuan, and it received another financing of over 100 million yuan in 2024, with a total of more than 600 million yuan. Panda Capital, Meijin Energy, GSR Ventures and other institutions are all among the investors.

In January 2021, ByteDance, Meituan-affiliated Kuxun Travel and angel investor Wang Gang entered the shareholding list, and the valuation at that time had reached USD 500 million.

However, in the early stage of Yinhe's development, expansion was already on the agenda.

On March 24, 2021, Yinhe signed a contract in Shunde for its global R&D and industrial application demonstration base, with a planned total investment of 10 billion yuan, covering an area of about 500 mu, located in Beijiao Robot Town. During the Qingming Festival holiday that year, Shen Gang also led the team to Shunde to negotiate the demonstration line. Nanfang Daily reported that the intended plot was still to be sold at that time. To promote the business as soon as possible, he planned to rent existing factories first to start the construction of the demonstration line.

Until 2024, Yinhe was still expanding cooperation. In July of that year, Sait Smart released official news showing that the two sides signed a strategic cooperation agreement. The directions introduced by Shen Gang included general-purpose robots in intelligent manufacturing, overall solutions, and smart city scenarios.

02 An intelligent manufacturing contract made the operational difficulties concrete

The public financial reports of Foshan Lighting have left a complete trail of Yinhe's journey from undertaking projects to litigation and refund.

In 2021, the two sides signed the "Construction Contract for the Intelligent Manufacturing Workshop Project of Bulb Lamps". According to the disclosure in Foshan Lighting's 2023 annual report, the project failed to meet the final acceptance standards, and the two sides failed to reach an agreement after negotiation. In December 2023, Yinhe sued Foshan Lighting, demanding payment of the contract price, interest, refund of the performance bond and compensation for losses, totaling about 104.4 million yuan.

Foshan Lighting then filed a counterclaim, arguing that the project had not met the final acceptance standards after overdue, and demanded the return of the paid contract price and the payment of liquidated damages.

According to the disclosure in Foshan Lighting's 2025 annual report, the first-instance judgment in September 2024 rejected all Yinhe's claims, ordering it to return 26.904 million yuan and dismantle the equipment and systems in the workshop at its own expense. In February 2025, the second instance upheld the original judgment. After entering the execution stage, the relevant equipment was evaluated at 49.7403 million yuan, and the first auction in April 2026 failed to attract any bids.

Originally hoping to recover the payment through litigation, it finally had to refund and dismantle the equipment. This reversal will bring direct capital pressure to an enterprise that still needs continuous investment in R&D and operation.

This single contract alone cannot explain all of Yinhe's losses and debts, but it at least reveals a verifiable operational setback that has already happened: the company got the project, but failed to smoothly complete the conversion from project to revenue and then to cash.

The difficulties of the robotics business are precisely concentrated in this process.

For factory customers, purchasing an automated system requires it to meet the agreed cycle, yield, stability and maintenance requirements in actual production. The equipment can run in the demonstration, and the entire production line can meet the acceptance conditions for a long time, there may be a lot of debugging work in between.

These tasks require engineers to be stationed on site, repeated modification of software and hardware, and coordination of original equipment, processes and production management. The project cannot be accepted for a long time, but the expenses of personnel and suppliers will not stop at the same time. If the contract also involves refund liability, the R&D and delivery costs that the enterprise has invested may not be recovered.

Therefore, from the perspective of order delivery of robot enterprises, delivery conditions often determine the actual value of an order to the company more than the figures at the time of signing.

Another layer of controversy surrounding Yinhe lies in corporate governance and the use of funds.

In June this year, Li Lun publicly called out to He Jianfeng, raising questions about executive compensation, office decoration expenses, project decision-making and shareholder exit arrangements. These accusations brought Yinhe's operational crisis into the public eye, but the relevant amount and ownership of responsibility still need to be distinguished from audits, transaction documents and judicial determinations. As of the release of relevant reports, the media had not obtained a public response from the Infore side.

The questions it raises are still worth discussing: when an enterprise has not yet generated stable revenue, who will restrain its expenditures? Who will judge whether a new project is worth taking? And who can call a halt to overly risky expansion in time?

Industrial background can help enterprises obtain resources, but it will not automatically establish effective budget, contract review and operational supervision. The inclusion of well-known institutions in the list of shareholders does not mean that they have the same degree of control over the company's daily operations.

The current public information is not sufficient to restore the entire capital flow of Yinhe. The expression of "burning up 600 million yuan" is still a simple connection between the financing scale and the end of the enterprise. To explain why it has reached the bankruptcy procedure, we still need to look at how the funds actually enter and are used, as well as the real relationship between project losses, operating expenses and debts.

03 Beyond Yinhe, other bankrupt robot enterprises

Two domestic cases that can be confirmed from judicial announcements are Deep Blue Robotics (Shanghai) Co., Ltd. and CloudMinds Robotics (Beijing) Co., Ltd.

The Shanghai Pudong New Area People's Court accepted the bankruptcy liquidation application of Deep Blue Robotics (Shanghai) on May 7, 2026, and the administrator later announced the termination of the labor relations of all on-the-job employees.

When the fifth batch of employee creditor's rights was publicly announced in September, the first four batches had confirmed 102 employees with total creditor's rights of more than 23 million yuan, including items such as wages, economic compensation, and provident fund. This amount cannot be simply summed up as "back pay", but it makes the cost of enterprise operation failure clear: in addition to capital losses, there are employees' labor remuneration and compensation that have been waiting for settlement for a long time.

The liquidation application for CloudMinds Robotics (Beijing) was accepted by the Beijing Fangshan District People's Court on October 10, 2025, and the administrator was appointed in June 2026, followed by the launch of creditor's rights declaration.

Looking overseas, there is no shortage of bankrupt robot enterprises.

The sweeping robot enterprise iRobot has a long history of commercialization, but it has also suffered severe cash pressure. Its financial report for the third quarter of 2025 showed that quarterly revenue dropped from USD 193.4 million in the same period of the previous year to USD 145.8 million; cash and cash equivalents dropped from USD 40.6 million at the end of June that year to USD 24.8 million at the end of September. The company cited market pressure, production delays and shipping disruptions, and stated that there were no additional capital sources available for withdrawal at that time. On December 14, 2025, iRobot filed for protection under Chapter 11 of the U.S. Bankruptcy Code. But it did not stop all business at that point: on January 23, 2026, the company announced the completion of a court-supervised transaction, exited the pre-packaged reorganization, and Picea acquired all its equity to continue operations.

The experience of German robot enterprise Franka Emika presents another kind of difficulty. In August 2023, the company entered temporary bankruptcy administration. It explained in the announcement at that time that the differences between shareholders prevented the completion of financing negotiations with potential investors, and emphasized that the company still had orders. Later, its business was acquired by Agile Robots and continued under the name of Franka Robotics.

04 These failures occurred in a booming market

Data released by the International Federation of Robotics (IFR) in September 2026 shows that in 2025, the global new installations of industrial robots exceeded 600,000 units, a year-on-year increase of 11%; China installed 354,000 units, a year-on-year increase of 20%, accounting for 59% of the global installation volume. In the same year, the installation volume in the United States increased, while that in Japan and Germany declined.

Factories are still buying robots, and the overall market is maintaining rapid growth.

However, this set of data only counts industrial robots, which cannot prove that the business models of humanoid robots, commercial service robots or consumer robots have matured. The customers, selling prices, delivery methods and after-sales costs of different products vary greatly, and the successful experience of one enterprise may not be easily replicated to another type of robot.

The tracks that Yinhe initially laid out were emergency management, urban governance and intelligent manufacturing. If its predicament is simply attributed to "the failure of the humanoid robot boom", the timeline and business background will be misaligned. But for the new round of entrepreneurs, it is still a sample worth disassembling in advance: technical capabilities, order scale and operational capabilities must grow synchronously in the same company, and none of them can be missing.

It is especially necessary to be alert to a trap hidden by the heat of financing: the enterprise keeps winning new projects, and the team size continues to expand, but every time it enters a new customer, it has to re-invest a lot of customized development and on-site delivery. The revenue on the books is rising, but the delivery cost may not decrease, and the capital occupation is getting heavier and heavier.

At this stage, what the company has to answer is no longer "how big can we be", but a series of extremely specific operational questions: which modules can be reused, which requirements should be decisively rejected, how many devices one engineer can support, how long a cycle a project takes from signing to payment collection, and how much loss the company can still bear after a quality dispute arises.

The real scale advantage that robot enterprises are worth accumulating should be reflected in: more stable subsequent delivery, shorter debugging cycle, lower maintenance cost, and customers' willingness to repurchase. Expanding the team and adding product lines can only truly improve operations if they bring the above changes.

What also needs to be iterated is the external coordinate system for evaluating a robotics company.

Investors can continue to pay attention to the team, technology and market space, but for companies that have entered the delivery stage, the acceptance rate, payment collection cycle, after-sales cost, customer repurchase rate and capital reserve deserve higher weight. The reputation of industrial shareholders cannot replace the penetrating verification of project quality.

Customers also need to incorporate the supplier's sustainable operation capability into procurement decisions. Robot equipment is often accompanied by long-term maintenance, spare parts and software support. After the equipment is delivered to the site, whether the supplier can continue to perform the contract directly affects the customer's use cost over the entire life cycle.

For Yinhe, the most worth tracking materials next are the progress of property investigation, creditor's rights verification and asset disposal in the bankruptcy procedure. They will further answer: how much realizable assets this company has left, which debts can be repaid, and what consequences the past business decisions have finally formed.

In those years, investors discussed how big a company it could grow into; now, creditors need to confirm how much of each sum of money can be recovered.

Between the two sets of calculations, there are R&D, manufacturing, debugging, acceptance and payment collection. Yinhe's experience has pushed these links that were once hidden behind the financing news back to the spotlight of the entire industry.