Li Zexiang's fourth IPO has yielded a 400-fold return

Recently, Benmo Power (Beijing) Co., Ltd. submitted its prospectus to the Hong Kong Stock Exchange, planning to list on the Hong Kong Main Board. The offering is underwritten exclusively by CITIC Securities (Hong Kong), with a planned issuance of no more than approximately 122 million H shares.

Benmo Power is a company that develops direct-drive force modules, the so-called robot joints. Its shipment volume reached 8.5 million units in 2025, capturing a 61.1% market share in the segmented market of direct-drive force modules for consumer robots in China, firmly taking the top position. The founder, Zhang Di, is only 31 years old this year. Zhang Di graduated from Beijing Institute of Technology with a bachelor's degree, then pursued a master's degree in robotic systems and control engineering at the Hong Kong University of Science and Technology. His supervisor is none other than Li Zexiang, widely known in the venture capital circle as the "Godfather of Robotics".

The entrepreneurial journey of Benmo Power has almost strictly followed the standardized process of Li Zexiang's entrepreneurship incubation model:

It identified the technical direction in the laboratory, moved to Songshan Lake to start the business, received the first round of funding from Li Zexiang right after the company was founded, then completed engineering verification and mass production in Dongguan, got subsequent investments from mainstream VC institutions, and finally moved towards the Hong Kong Stock Exchange.

If Benmo Power is successfully listed, it will become the fourth IPO nurtured by Li Zexiang, following Googol Technology, Idriverplus, and Woan Robotics.

Li Zexiang has cultivated another post-90s entrepreneur

In March 2020, Benmo Technology was registered and established at the Songshan Lake International Robotics Industry Base in Dongguan, with the startup capital of RMB 300,000 invested by Li Zexiang's Songshan Lake Robotics Institute. At that time, the founder Zhang Di was only 26 years old, having just finished his master's degree.

The actual investment process is even more legendary. According to Zhang Di's account, he only went to Li Zexiang to propose an initial idea at first, and Li Zexiang simply told him to "go to XbotPark to get hundreds of thousands of yuan". Zhang Di was confused and asked what the money was for, and Li Zexiang replied: "Use it as travel expenses first, and investigate whether this direction is worth exploring." XbotPark is the official English name of Songshan Lake International Robotics Industry Base.

The idea Zhang Di put forward can be summed up in one sentence: "Eliminate the reducer". In traditional robots, no matter what shape they are, the reducer is an indispensable transmission device. However, the reducer has two defects: first, high-precision micro reducers are extremely difficult to manufacture, and this is also a "stuck neck" technical field; second, the reducer will cause noise and mechanical wear.

If you want to eliminate the reducer, you need to adopt a direct-drive solution, that is, let the motor directly drive the robot joints without relying on transmission components.

Of course, direct drive is not a new concept, and it has long been applied in fields such as high-precision machine tools and advanced semiconductor equipment. However, due to factors such as torque density, energy consumption, volume and cost, direct-drive solutions have not been popularized in robots for a long time. What Benmo Technology does is to solve these problems through engineering means.

For example, Benmo Technology abandoned the bulky traditional design of "external driver + discrete motor", and highly integrated the high-resolution magnetic encoder, temperature sensor array, inverter power bridge and MCU controller on a single board inside the motor end cover, which reduces the axial thickness of the motor by about 50% and shrinks the volume by one third. By connecting the thermal interfaces between the winding, stator core, bearing and shell, the heat of the motor is directly conducted and dissipated through the extended end of the rotating shaft and the metal end cover, thus achieving a torque density as high as 15~20 Nm/kg, and solving the problems of overheating demagnetization and motor burnout under closed working conditions.

Benmo Technology itself also manufactures robots. In 2021, Benmo Power launched the direct-drive two-wheeled legged robot "Xingtian", and in 2023, it launched the 8-degree-of-freedom two-wheeled legged robot TITA.

But what really pushed Benmo Technology to the listing stage is the external sales of direct-drive modules. The prospectus shows that from 2023 to 2025, the revenue of Benmo Technology's robot power module business increased from RMB 14.217 million to RMB 282 million, growing 20 times in two years, and its proportion in total revenue also rose from 81.1% to 97.4%. In the field of consumer direct-drive modules, Benmo is the absolute leader in the industry. In 2025, Benmo Technology ranked first in China's consumer robot direct-drive force module industry, with a market share of 61.1%. Its customers cover four of the world's top ten consumer robot providers.

The so-called consumer robots mainly refer to household cleaning robots, such as sweeping robots, washing and mopping all-in-one machines, as well as lawn mowing robots, AI companion robots, etc.

While its performance continues to grow, the valuation of Benmo Technology has also exploded. From 2020 to 2025, Benmo Technology completed 12 rounds of financing, with total financing exceeding RMB 650 million. Well-known institutions such as 5Y Capital, Lenovo Capital, Songhe Capital, Legend Capital, as well as industrial capitals such as XPeng and SenseTime have all entered the market. By the time of the Series C financing in December 2025, the post-investment valuation of Benmo Technology had reached RMB 3.185 billion.

In 2020, Li Zexiang's RMB 300,000 seed round investment only had a post-investment valuation of RMB 5 million. After considering subsequent share dilution, it still achieved a 400x return.

The "Godfather" who understands young entrepreneurs best

The success of Benmo Technology has once again proved one thing from Li Zexiang's perspective: a fledgling young person with strong engineering capabilities, if correctly identified and provided with laboratory resources, capital and enough time, can truly be "cultivated" into an entrepreneur.

As a professor, Li Zexiang does not focus on grades when recruiting students, and almost all of his students are people who love to explore and try new things. The most widely talked about case is that Wang Tao only got a C grade for his graduation design, but Li Zexiang admitted him through special approval because of his excellent performance in the Robocon competition. Zhang Di is also such a person who loves to explore. When he was a student, he developed a set of aerospace attitude controllers, and finally sold the whole set of products and technology to a satellite company, earning hundreds of thousands of yuan.

In 2004, Li Zexiang introduced the national college student robot competition Robocon into the classrooms of HKUST, and set up the course "Robot Competition and Design", allowing students to form teams, select topics independently, and complete robot R&D and competition within 8 months. This course deliberately abandoned the credit-only assessment rule, and evaluated students' multidisciplinary cross-capability, practical ability and supply chain management ability - the latter two are exactly the capabilities that cannot be taught in traditional classrooms. Wang Tao was the beneficiary of the first session. In 2004 and 2005, he led the HKUST team to win the Hong Kong regional championship of Robocon and the third place in the international competition successively.

After that, Li Zexiang began to systematically recruit "disciples" from the participating teams of various universities, forming a selection funnel of "selecting talents through competitions". Tao Shizheng, the founder of ePropulsion, once served as the captain of the HKUST robot team and participated in the Asia-Pacific College Student Robot Competition for many times. Zhang Junbin, the founder of Narwal, also won Li Zexiang's favor with his past competition experience in the Feibao Cup and HFR team, and became the first project talent introduced into the Songshan Lake Robotics Industry Base.

Twenty years later, from the 3126 Laboratory of HKUST, to XbotPark in Songshan Lake and the Qingshui Bay series of funds, Li Zexiang has built a unique entrepreneurship system for young people, and he himself has become the person best at creating "hit" projects in the robotics field.

Of course, the reason why Li Zexiang's students/disciples have such a high success rate is far more than Li Zexiang's ability to select talents. In fact, a robotics entrepreneurship ecosystem that is probably unique in China has been formed around Li Zexiang.

For example, Googol Technology, founded by Li Zexiang himself, is China's first motion controller enterprise with independent intellectual property rights, which provides underlying motion control technology for other companies in the system. When DJI conquered the brushless direct-drive gimbal in 2011, it drew on the technical accumulation of Googol in the field of CNC machine tools.

After DJI's success, it pushed Li Zexiang's system to a new level. One of the most difficult hurdles for entrepreneurship in the robotics track is the supply chain, especially for young people who have just stepped out of campus. In the early days of DJI's entrepreneurship, Li Zexiang personally took Wang Tao to Huaqiangbei to teach him how to bargain and identify electronic components. After DJI's success, it became the "supply chain parent" of Li Zexiang's system, opening up supply chain resources to subsequent startup teams.

4 IPOs, 12 Unicorns

In August 2023, Googol Technology was listed on the ChiNext board. On December 19, 2025, Idriverplus, which Li Zexiang co-founded and served as chairman, was listed on the Hong Kong Stock Exchange. 11 days later, on December 30, Woan Robotics incubated by XbotPark went public.

Now, Benmo Power is also approaching the listing stage. If it is successfully listed, the number of publicly verifiable listed companies under Li Zexiang's system will increase to 4. In addition, HAIROBOTICS has submitted its listing application, and Narwal is also in the Pre-IPO stage.

The closed loop of "discover talents through competitions — cultivate talents in the laboratory — incubate projects in the base — invest the first sum of money through funds — complete mass production with the Pearl River Delta supply chain — obtain follow-up financing from VCs — exit through the capital market" has been fully validated by Li Zexiang. According to an exclusive interview with Li Zexiang published on August 10, as of 2025, this system has incubated more than 280 hard technology enterprises in total, with a total valuation of over RMB 500 billion, of which 4 are listed and 12 have grown into unicorns or quasi-unicorns.

Unicorn star companies not yet listed under the "Li Zexiang System"

Li Zexiang basically does not follow the mainstream VC model of "heavy investment in multiple rounds". He usually takes 5%-13% of the shares in the early stage, and basically does not add positions after the Series A round, leaving enough space for mainstream VCs, and his shareholding will be diluted to 0.2%-6% before IPO. For example, before Benmo Technology goes public, Li Zexiang's total shareholding after penetration is only 3.39%.

Therefore, this is a typical strategy that pursues odds rather than winning rate. As long as more than a dozen unicorns emerge from 280 enterprises, the overall IRR of the system can be realized. In Li Zexiang's words: "It's like walking on a dark road, one person will feel scared, but if several people walk together, and you are given a flashlight and a stick to fend off dogs, you will be much safer."

From another perspective, it can be said that Li Zexiang has spent more than 20 years turning the originally highly accidental seed-stage investment into a "assembly line" model that is expected to be stably replicated. After achieving success in Songshan Lake, Li Zexiang is also replicating this system to Ningbo, Changzhou, Chongqing, Hunan and other places. Of course, these new bases cannot replicate the talent success density of Songshan Lake in the short term, and the "replicability" of the model still needs to be fully verified. If Li Zexiang succeeds, the primary market will never lack high-quality investment targets again.

This article is from the WeChat official account "ChinaVenture", written by Tao Huidong, and published by 36Kr with authorization.