12.5 billion yuan, Xiong'an ushers in its first IPO on the Hong Kong Stock Exchange

On September 1, Mech-Mind officially listed on the Hong Kong Stock Exchange.

For institutional shareholders that have followed this company for four to six years, this is undoubtedly a moment of harvest.

HSG has invested in four consecutive rounds since 2020 and remains the largest institutional shareholder of the company before its listing, holding 13.63% of the shares; Source Code Capital first made its move in 2020, co-led the Series B+ round and continuously followed up the Series C and Series C+ rounds, accompanying the company for nearly six years; Meituan, IDG Capital, Qiming Venture Partners, Intel Capital... a so-called "top-tier" investor list is finally about to deliver the answer to its ten-year layout on the Hong Kong stock market.

However, if this is only regarded as a financial victory, the story would be far too one-sided.

Hidden between the lines of the prospectus is another more thought-provoking beneficiary — Xiong'an New Area. In 2024, Mech-Mind moved its global headquarters from Beijing to Zhongguancun Science Park in Xiong'an New Area.

A unicorn whose products are sold in nearly 50 countries has rooted itself in a future city that is still under construction. When the bell rings, the harvest belongs not only to the investors in the audience, but also to the heavy transcript that this young city has obtained for the first time.

One is a ten-year entrepreneurial story that has finally paid off after years of hard work, one is the victory of capital long-termism, and one is the industrial footnote of a new area rising quietly — the three lines intersect and weave into the true weight of Mech-Mind's listing this time.

From a Match Result of AlphaGo to Ten Years of Dedication in Factory Workshops

On March 9, 2016, Shao Tianlan, a Go enthusiast, watched the fourth game of AlphaGo against Lee Sedol on the screen.

4 to 1, the machine won.

Recalling his feelings at that time many years later, he described it as "adrenaline rush" — not because of the result of the Go game itself, but because he suddenly realized that the field of artificial intelligence might really be about to usher in a game-changing era.

He was 27 years old that year, having just obtained a master's degree in robotics from the Technical University of Munich, and working as an engineer at a well-known collaborative robotics enterprise in Germany.

According to most people's life paths, the next step would be to stay in Europe, or return to China with his resume to join a large factory. In fact, most of his classmates from the School of Software at Tsinghua University went to Google, Microsoft, Huawei, and BAT. "Almost no one like me plunged into the industrial field." That AlphaGo match became a "trigger".

Seven months later, he declined offers from many Silicon Valley companies, and founded Mech-Mind together with two Tsinghua alumni Fu Ao and Ding Youshuang, plunging headlong into the most unglamorous scenarios: factories and logistics warehouses.

"Industrial robots sound extremely impressive, but in reality they are clumsy and difficult to use, far from the ideal state of human helpers." Shao Tianlan later explained what they were doing: to truly install "intelligence" into those robotic arms that repeat labor on the assembly line every day.

Shao Tianlan has a widely spread metaphor for how to take this path: starting a business is like driving a car rushing towards a cliff. The competition is not about who drives faster, but about who can get closest to the cliff without falling off. The control of the accelerator and brake is more important than anything else.

Over the years, the robotics and embodied intelligence track has gone through several rounds of highlights and troughs, and too many companies have been seen either "stepping on the accelerator too hard" and dying halfway, or "stepping on the brake too early" and failing to break through the siege.

Mech-Mind also faced tight capital during the industry trough, but it did not chase high valuations when the market was the hottest, nor did it slam on the brakes when the market was the coldest. Instead, it has always focused on whether the products can be truly implemented.

An investor recalled that in the early days, in order to promote project implementation, Shao Tianlan even lived in the factory. He rushed back to Shanghai from a remote factory area overnight for a meeting, and had to go to the bathroom to wash his face and shave before daring to sit in the conference room — this almost "rooted in the soil" entrepreneurial style is precisely what distinguishes Mech-Mind from many "laboratory-type" AI companies.

Ten years later, this team that regards "being able to get work done" as the highest pursuit has delivered a satisfactory report card: its products have been deployed on more than 27,000 units worldwide, covering dozens of industries and hundreds of segmented scenarios. Calculated based on 2025 revenue, Mech-Mind ranks first in the global market of AI+3D vision guided general intelligent robot components, with a market share of about 22.1%, and also ranks first in the world in terms of shipment volume.

What is more thought-provoking is its positioning choice.

At a time when embodied intelligence is wildly sought after by capital and almost all players are betting on "making complete machines and humanoid robots", Mech-Mind deliberately did not make a complete robot, but focused on making the "eyes, brain and hands" of robots — the Mech-Eye industrial 3D camera is responsible for "seeing", with accuracy ranging from 0.03 to 0.2 mm, far exceeding the industry's average level of 0.5 to 1.0 mm; the self-developed Mech-GPT multimodal large model is responsible for "thinking", with recognition accuracy exceeding 99.99%; the Mech-Hand five-finger dexterous hand is responsible for "doing", completing fine grasping. The three parts together form a complete closed loop from perception, decision-making to execution, making Mech-Mind a typical "shovels seller" in this round of embodied intelligence wave — no matter who finally manufactures the complete machine, it is very likely to need its "eyes, brain and hands".

This is a slower path that is harder to be directly understood by the capital market, but after ten years of walking, it is precisely this "slowness" that has brought it an unshakable first market share in this segmented track.

From rejecting the Silicon Valley offer and returning to China at the age of 27, to bringing the company to the gate of the Hong Kong Stock Exchange today, Shao Tianlan has spent nearly ten years proving one thing: real technological breakthroughs never happen overnight, but depend on whether a person is willing to stay in the factory workshop and push something that sounds not so glamorous to the extreme.

Wager the Capital Before the Consensus is Formed

If the entrepreneur's story is a ten-year long run of a lone warrior, then the investor's story is about whether a group of people are willing to put trust and resources in advance at the stage when "the answer is unclear".

Gong Yuan, Partner at HSG, still remembers the experience before the investment decision in 2020. At that time, the robotics industry was generally in a downturn, "embodied intelligence" was not even a commonly used term, and the AI capability of hand-eye coordination just showed a little sign, far from the stage of large-scale shipment. For most investors, robots were still equivalent to industrial automation robotic arms running according to rules, and no one had really seen how a robotic arm with generalization capability could independently complete tasks in different scenarios.

To figure out whether this matter is reliable, a team of three investors from HSG spent two months flying to ten cities, visiting steel mills, logistics factories, clothing factories and automobile factories where Mech-Mind was carrying out pilot implementation at that time one by one, and made dozens of expert interview calls.

After the round of visits, they came to a clear judgment: Mech-Mind already ranked first in market share, and was accumulating real reputation and brand recognition. With these first-hand materials accumulated from on-site inspections, the HSG team demonstrated to the investment committee how AI enables a robotic arm to obtain generalization capabilities, finally completed the investment, and then added positions three times in a row.

Compared with the investment logic, Gong Yuan prefers to talk about a detail related to people.

He and Shao Tianlan are the same age, and have known each other for nine years. "Our relationship is not just that of investors and entrepreneurs, but more like partners growing up together."

One morning, the two arranged to meet at the HSG office in Shanghai Plaza 66. Shao Tianlan showed up with breakfast in his left hand, a toiletries bag in his right hand, and a suitcase in tow — because he had just rushed back to Shanghai from a factory in another city overnight the previous night. He had been living in the factory for days to promote product implementation, looking unkempt, so he could only go to the bathroom to wash his face and shave before sitting in the conference room.

After that, Gong Yuan simply stopped arranging meetings with him in office buildings, and tried to choose places closer to the factory to meet. Behind this episode is a rare trait: a returnee AI entrepreneur is willing to squat in the factory instead of sitting in the office drawing blueprints.

If HSG's story is "visiting all factories before placing a heavy bet", then the story of Source Code Capital is about another kind of patience.

Source Code Capital first invested in Mech-Mind in 2020, co-led the Series B+ round, and then successively participated in the Series C and Series C+ rounds, investing a total of three rounds before and after, accompanying the company for nearly six years, and is still one of the important institutional shareholders of the company before listing.

It took Mech-Mind ten years from its establishment to entering the capital market, and Source Code Capital accompanied it for nearly six years — putting these two time spans together, it exactly conveys a simple truth: a direction that has not yet become the industry consensus requires someone to believe in it first, and someone who is willing to spend enough time to accompany it to be verified.

What Source Code Capital values is never how much revenue Mech-Mind achieved at that time, but several longer-term questions: Will the combination of AI and robots become a clear industrial direction? Can standardized products really be copied across industries and different robot bodies? Has the team really figured out what to do and what not to do? After the first investment, Source Code Capital successively participated in two more rounds of financing — each additional investment is a re-verification of whether the company's direction remains stable, and whether customer needs have moved from exploratory attempts to real production and large-scale replication.

HSG's on-site investigation across ten cities, Source Code Capital's six years of companionship and unrestricted patience, plus a luxurious shareholder list spanning financial investors and industrial capitals such as Meituan, IDG Capital, Qiming Venture Partners, Intel Capital — these institutions have already put real money and long-term trust into Mech-Mind when it was still a "non-consensus" story.

At the moment of the bell-ringing ceremony at the Hong Kong Stock Exchange, what is realized is not just an investment return, but also the judgment of "willing to bet in advance before the answer appears", which finally waits for its own moment of verification.

Past Events of Xiong'an: A Future City That Holds an Industry Facing the Future

If the stories of entrepreneurs and investors talk about how "people" make unwavering choices for ten years, then the story between Mech-Mind and Xiong'an is about the mutual achievement between a city and an enterprise.

In November 2024, Mech-Mind's global headquarters located in Zhongguancun Science Park, Xiong'an New Area was officially inaugurated. Talking about why he moved the headquarters from Beijing to here, Shao Tianlan used a word — "two-way rush".

"We are engaged in an industry facing the future, and Xiong'an is a full-of-futurism city." He explained that Xiong'an New Area's complete artificial intelligence industrial planning, rich scientific and technological innovation resource elements, talent-friendly policies and excellent business environment have provided sufficient soil for the development of Mech-Mind.

This is not just a polite remark.

After settling in Xiong'an, the new area has implemented "full support" through a complete set of specific and detailed service systems.

Talent is the first step: shortly after the headquarters was inaugurated, the relevant departments of the new area took the initiative to visit the enterprise, and explained in detail the talent policy known by employees as "16 Talent Policies for Xiong'an"; the new area provides subsidies for new graduates who join the enterprise at the standard of 1000 yuan per month for undergraduates, 2000 yuan per month for masters, and 3000 yuan per month for doctors.

Shao Tianlan said that such dividends make young talents feel the care in practice, and further stimulate the team's innovation enthusiasm. Living facilities are another attraction — it only takes more than ten minutes to commute between the workplace and the residence every day, with numerous parks and green spaces, libraries, sports centers, and commercial streets all available. "After coming to Xiong'an, I obviously feel that everyone has more smiles on their faces and more motivation at work."

What is more critical than talent policies and living facilities is the innovation soil brought by industry-university-research collaboration.

In April this year, with the support of the new area, Mech-Mind and Xiong'an Aerospace Information Research Institute jointly built an embodied intelligence R&D center; the enterprise also carried out collaborative innovation with Tsinghua University, Hebei University of Technology and other universities in cutting-edge directions such as embodied intelligence large model and physical AI, so that technological achievements in the laboratory can be put into production lines faster. The aggregation effect of industrial ecology is also happening simultaneously: with the help of the new area, Mech-Mind, together with Xiong'an Group, iFlytek and other enterprises, jointly initiated the establishment of Xiong'an New Area Zhongguancun Robot Industry Alliance, where upstream and downstream enterprises closely cooperate in key technology collaboration and scenario expansion.

"The robot industry has a long chain and covers a wide range of disciplines, and everyone especially needs to work together." Shao Tianlan said that after coming to Xiong'an, the enterprise quickly integrated into a vibrant industrial ecosystem.

This confidence was quickly transformed into tangible business growth.

After settling in Xiong'an, Mech-Mind accelerated the improvement of the general embodied intelligence "brain-eye-hand" physical AI collaborative system — the Mech-GPT large model as the "brain" deeply cooperates with the industrial camera "eye" and the dexterous manipulator, upgrading the robot from programmed mechanical execution to more intelligent autonomous operation. In the field of automobile manufacturing, this system is applied to the whole process of stamping, welding, assembly and quality inspection; in the field of logistics warehousing, the efficiency of palletizing, sorting, loading and unloading has been significantly improved; it has also achieved large-scale implementation in new energy, consumer electronics, food and medicine and other fields.

While the industrial market is advancing rapidly, the door to the civilian market has also been quietly opened. In the exhibition hall of Xiong'an Artificial Intelligence Industrial Park, the staff only need to "place an order" on the tablet, and a humanoid robot can flexibly open its hand, accurately pick up a can of Coke from the shelf and hand it over; at another booth, the staff says to the robot equipped with the Mech-GPT large model "Please put the food that monkeys eat into the box", and the robot can accurately identify and grab the banana from models such as carrots, green peppers, and bananas — behind these seemingly light demonstrations is the technical closed loop that can only be supported by the deep collaboration of "eye", "brain" and "hand".

On June 25 this year, Hurun Research Institute released the "2026 Global Unicorn Index". Among the 1603 listed enterprises worldwide, Mech-Mind from Xiong'an New Area is particularly eye-catching — this is the first time that this embodied intelligence enterprise founded by the Tsinghua returnee team has appeared on a global industrial list with the identity of "Xiong'an".

Ten years ago, Shao Tianlan saw the turning point of an era in the result of AlphaGo's match; four years ago, HSG's investors flew to factories across ten cities to verify a judgment that had not yet taken shape; two years ago, Xiong'an New Area opened its arms to this enterprise.

The three timelines finally converged in this summer of 2026 — a robotics company that grew out of factory workshops is about to ring the bell belonging to itself and this future city at the Hong Kong Stock Exchange.

This article is from the WeChat Official Account "Rongzhong Finance" (ID: thecapital), written by Abu, edited by Wuren, published with authorization from 36Kr.