Suzhou has seen another unicorn company go public.
On July 8, 2026, a physics AI unicorn project debuted on the Main Board of the Hong Kong Stock Exchange.
Momenta, a Suzhou-headquartered physics AI company, entered the capital market with an IPO price of HK$295.6, raising an estimated HK$58.9 billion. Fourteen cornerstone investors subscribed a total of HK$30 billion, with China International Capital Corporation and Deutsche Bank acting as joint sponsors. The offering adopted no price range and locked in the final pricing directly — such a level of issuance enthusiasm is rarely seen among Hong Kong stock IPOs in recent years.
As of now, Momenta's market capitalization stands at HK$700 billion.
More compelling than the capital figures is a decade-long story, tied to one street, about how an engineer turned an ideal into tangible engineering reality.
Notably, this physics AI unicorn rising from the autonomous driving track hails from Suzhou. In the first half of this year, Suzhou maintained a strong momentum, with projects such as Linkin Instruments and Gude New Materials completing their IPOs one after another. Momenta's successful listing now adds another prominent name to Suzhou's capital market landscape.
A Decade Ignited by One Street
One afternoon in 2015, in Mountain View, Silicon Valley.
Cao Xudong drove past an unassuming street — Fairchild Drive. This name means nothing to most people, but for a technologist researching computer vision, it carries immense weight. It was on this street that Fairchild Semiconductor was born. The rebellious group of "Traitorous Eight" founded the company, from which dozens of semiconductor giants including Intel and AMD later emerged, directly giving rise to the name "Silicon Valley."
Standing on that street, the spark of ambition in Cao Xudong's heart was ignited.
At that time, Cao Xudong was already the head of Beijing R&D at SenseTime, a Tsinghua University graduate, and an alumnus of Microsoft Research Asia, with years of deep experience in the field of computer vision. He lacked nothing — technical expertise, professional prestige, and a high salary.
But Fairchild Drive showed him another possibility: Could a company define the future of autonomous driving in the same way Fairchild Semiconductor defined the chip industry?
In September 2016, Momenta was founded.
The formation of the founding team of a startup relied on reputation, proven achievements, and a blueprint whose full picture was not yet clear. Core members came from the early SenseTime team, Tsinghua University, and the University of Science and Technology of China — a group of people who had won championships in the algorithm world, determined to translate what was written in academic papers into real vehicles. The 8 founding shareholders were all core members who joined in 2016, each holding equity incentives, with their eyes set on a seemingly unreachable goal.
But Cao Xudong soon discovered a harsh reality.
To achieve fully driverless operation, approximately 100 billion kilometers of road test data is required. What does that mean? It is equivalent to millions of vehicles driving on the roads for a full year. For a startup, this cost was simply unthinkable. Back then, the industry consensus was to pursue Level 4 autonomy — directly building driverless cars, running demos, conducting tests, and crafting compelling narratives. Almost all competitors were taking this path.
Cao Xudong deliberately chose a path that was "not glamorous enough."
He set the strategy of "one flywheel, two legs" for Momenta: One leg focuses on mass-produced Level 2 assisted driving. Through partnerships with automakers, mass-produced vehicles equipped with Momenta's systems travel across streets and lanes, generating massive amounts of data. The other leg focuses on Level 4 fully driverless technology, using this data to train algorithms and solve long-tail edge cases. The two legs move forward alternately: data from mass-produced vehicles trains driverless algorithms, and driverless technology in turn enhances the experience of mass-produced vehicles.
The advantage of this path is its clear logic, while the disadvantage is — it is extremely tough.
A research-oriented company had to humble itself to work closely with automakers, iterating on requirements, refining deliveries, and perfecting every engineering detail. Cao Xudong himself admitted that the process of mass-producing the first product "was quite challenging."
A friend familiar with Cao Xudong observed a detail: In the early years, his social media posts were lively, sharing daily work moments, family and friends, and his children's growth. Then for a period of time, all his posts were the company's PR articles. In the past year or two, he started sharing Huawei's management philosophy. "From a technical manager to a company leader, he has changed a lot over these years."
This transformation is the mark of a decade of tempering.
Looking back now, the data does not lie. The number of mass-produced vehicles equipped with Momenta's systems has surged from nearly 300,000 to over 1 million, an increase of more than 160%. The company has delivered more than 70 mass-produced vehicle models in total, with over 200 models designated for future deployment. In 2022, the first 100,000 units of mass production took 24 months, but now the fastest delivery of 100,000 units can be completed in less than 40 days. The flywheel effect is accelerating.
More critically, Momenta is no longer just an intelligent driving company. It is advancing Robotaxi tests in Munich, Germany in partnership with Uber, and exploring the Southeast Asian market with Grab. Its invested portfolio company, Lingyi Automotive, has entered the new energy intelligent heavy truck market, and Momenta itself has begun to build its own driverless truck team. From passenger vehicles to commercial vehicles, from China to the world, Momenta's business landscape is unfolding.
At the 2026 Beijing Auto Show, during the sharing session for Momenta's 10th anniversary, Cao Xudong recalled that afternoon on Fairchild Drive. The spark ignited ten years ago is now burning even brighter. He is working to join hands with all Chinese AI companies to jointly write a Silicon Valley-style legend belonging to the East.
The Person Who Decided to Issue the TS Immediately After the Meeting
Behind every great company, there are several people who placed the right bets at critical moments.
Wang Xiao, founder of Heji Venture Capital, still clearly remembers the scene of his first meeting with Cao Xudong.
In 2016, at the Continental Grand Hotel Beijing, Cao Xudong sat across from him, talking about his academic background at Tsinghua, computer vision research at Microsoft Research Asia, and practical experience at SenseTime. But what truly impressed Wang Xiao was not his resume, but his fundamental thinking about autonomous driving.
"He realized very early that the massive long-tail problems of Level 4 autonomous driving could never be solved with just a few hundred research vehicles — it must be deduced and captured through the scale of mass-produced vehicles," Wang Xiao recalled. During their first meeting, Cao Xudong described his vision of achieving full autonomous driving step by step across different levels, and repeatedly emphasized the importance of data.
"This aligned perfectly with Heji's view on data, which was a key factor leading us to make a decision very quickly."
Wang Xiao stated on the spot: We want to invest. "We should be the first institution to issue a term sheet (TS) to Momenta."
But the investment process was never smooth. The first round of investment did not go through for some reasons, which made Heji's colleagues quite regretful, believing that they "should have transferred the money directly back then." However, they persisted in following up and continued to invest in subsequent rounds. At that time, as a fund of modest size, Momenta's valuation in the first round had already exceeded their usual investment range. But because of their firm conviction, they allocated the largest single project investment amount in Heji's history.
"Although some people at the time thought this path was 'not glamorous enough,' the industry consensus was that Level 4 could be achieved directly, with technology and policies growing at a linear explosive pace, and almost all other startups were working on Level 4, demos, and test vehicles. But as a technologist and product person from the industry, I highly recognized the 'two-leg strategy' from day one," Wang Xiao stated frankly.
"Our non-consensus view was: Autonomous driving is not a simple software upgrade problem, but a long-term, data-scale-driven engineering ecosystem problem. Algorithm architecture, data accumulation, and integration into the automotive industry chain are all huge challenges."
Wang Xiao also admitted that his only worry was: "This path is so tough, can a research-oriented team endure the loneliness, humble themselves to collaborate and provide services to automakers? The results proved that they are more resilient than I imagined."
It is not just Heji.
Behind Momenta stands a long list of star-studded investors: BlueLake Capital, Sinovation Ventures, ZhenFund, Shunwei Capital, Yuanhe Capital, Gaoda Capital, Yunfeng Capital, Temasek, Toyota Motor, Mercedes-Benz China, GIC, Gaorong Ventures, Tencent Investment... From top-tier VCs to industrial capital, from sovereign funds to automotive giants, Momenta has secured investments from the most elite institutions.
The cornerstone investor lineup for this IPO is equally impressive — 14 institutions, with a total cornerstone subscription of approximately HK$30 billion. It includes top international long-term funds and sovereign funds focused on ultra-long-cycle value growth, top-tier strategic industrial investors, as well as leading Chinese long-term private equity, public funds, and insurance capital. Such a cornerstone lineup is truly rare in Hong Kong stock market initial public offerings in recent years.
Capital has cast its vote of confidence with real money. The underlying logic behind this trust is the commercial moat that Momenta has built over ten years.
Notably, Momenta's adjusted net loss has narrowed from 1.093 billion yuan in 2023 to 303 million yuan in 2025, with the loss-to-revenue ratio dropping all the way from 147% to 12.6%, approaching the critical point of operating break-even. From 2023 to 2025, its operating revenue grew from 743 million yuan to 2.413 billion yuan, with an annual compound growth rate of over 80%. In the current autonomous driving industry, where companies are generally facing heavy losses, this performance report is particularly outstanding.
As Wang Xiao said: "The market will always reward long-distance runners who do difficult but correct things."
Suzhou: The "Rainforest" for Unicorns
Momenta's headquarters is in Suzhou.
This city has once again stepped into the spotlight because of a company's IPO. But focusing only on Momenta would underestimate the true potential of Suzhou.
In the first half of 2026, Suzhou saw a total of 12 new domestic and overseas listed companies. In January, Ribio Biotech rang the bell on the Hong Kong Stock Exchange, becoming the first listed company in Jiangsu for the whole year. In the same month, Aishelun Technology debuted on the Beijing Stock Exchange, filling the gap in Xiangcheng District. In April, Linkin Instruments went public on the Sci-Tech Innovation Board, closing at 799 yuan on the listing day with an increase of 875.82%.
By the end of the first half of the year, Suzhou was home to 235 A-share listed companies and 47 Hong Kong-listed enterprises, with both the breadth and depth of its capital market landscape ranking among the top in China.
By the end of May, the total market capitalization of Suzhou's A-share listed companies exceeded 4 trillion yuan, ranking fourth among all cities in China. Seven companies — Dongshan Precision, TFC, Huadian, Hengtong Optic-Electric, Linkin Instruments, Nationalchip, and Robostar — have a market capitalization of over 100 billion yuan, while overseas listed company Innovent Biologics also surpassed the 100 billion yuan mark. In advantageous tracks such as optical communications, PCBs, and biomedicine, Suzhou has nurtured a group of "chain leader" enterprises that hold industry discourse power.
The cultivation of unicorns is equally remarkable. According to the Hurun Global Unicorn Index 2026, Suzhou ranks sixth in China with 16 unicorns, adding 4 new unicorns this year — the highest increment among cities excluding Beijing, Shanghai, and Shenzhen. These unicorns are distributed across multiple tracks including AI, biomedicine, new energy, semiconductors, and aerospace. They are not the product of a fleeting trend, but the result of simultaneous growth across multiple industrial chains.
A question worth exploring is: Why Suzhou?
Some say Suzhou benefits from the geographical advantage of "being close to Shanghai." But the distance between Suzhou and Shanghai has never changed. Before 2020, Suzhou was not particularly prominent on the unicorn rankings. The real turning point came when hard tech rose to prominence, and Suzhou's decades-old manufacturing foundation began to pay off.
When many places had not yet fully grasped the concept of guiding funds, Suzhou had already pioneered the model of "state-owned capital acting as LP, hiring market-oriented GPs for management, and then investing in projects." Suzhou poured funds into the early and mid-stages of integrated circuits, biomedicine, and advanced manufacturing. Later, Suzhou iterated to a new model of "chain leader + industrial fund" — state-owned capital does not act as an all-round player on its own, but works alongside already successful chain leader enterprises to supplement and strengthen the industrial chain, further reducing risks.
A strong county-level economy is another key factor. The four county-level cities under Suzhou — Kunshan, Zhangjiagang, Changshu, and Taicang — consistently rank among the top 10 top 100 counties in China, with economic volumes comparable to regular prefecture-level cities.
In addition, a cooperation network has formed between municipal and district/county-level funds — a project can be settled in Kunshan, with Suzhou's municipal government following up on the investment, and the industrial park providing talent policies and subsidies. This multi-level capital operation within a city, systematically operating at the prefecture-level city level, is envied by many other cities.
More importantly, it is Suzhou's understanding of "long-termism." The tracks where these unicorns are located — biomedicine, nanotechnology, photonics, semiconductors — share the common features of long cycles, heavy assets, slow returns, but high barriers to entry that are hard to replicate. The internet-style capital story of listing in three years and exiting in five years does not apply here. Suzhou's photonics industry has reached a scale of 360 billion yuan, achieving full-chain layout; Suzhou Industrial Park is one of the world's five major nanotechnology industry clusters; biomedicine is Suzhou's "No. 1 industry," with innovative pharmaceutical companies such as Innovent Biologics and Ascentage Pharma entering the capital market in batches.
Beyond Beijing, Shanghai, Shenzhen, Guangzhou, and Hangzhou, Suzhou has blazed a unique path: hard tech + patient industrial parks + market-oriented state-owned capital + county-level economic matrix. Without the traffic dividends of consumer internet, and not relying on central enterprise headquarters or ministry resources, Suzhou has spent more than 20 years making early layouts in multiple slow-burning tracks, attracting social capital by adhering to the principle of "state-owned capital acting as LP without arbitrary intervention," and expanding development space through several top 100 counties.
Suzhou has nurtured an ecosystem through services and capital, turning its decades-accumulated manufacturing foundation into fertile soil for new species to thrive.
On July 8, Momenta debuted on the Hong Kong Stock Exchange as the "first physics AI stock."
From that afternoon on Fairchild Drive to the bell-ringing moment at the Hong Kong Stock Exchange, Cao Xudong spent a full ten years. Over this decade, he transformed from a technical expert to a manager, from the leader of a group of algorithm engineers to the CEO with partnerships among the world's top 10 automotive groups. Momenta also grew from a startup consisting only of computer vision algorithm engineers to an industry giant that cooperates with 9 of the world's top 10 automotive groups, with business covering more than 10 countries and regions.
And Suzhou, the city that gave rise to Momenta, is continuing to write its industrial legend with a density of 12 new stocks in half a year.
The story of that street where Fairchild Semiconductor was located belongs to the 20th century Silicon Valley. The story of China's physics AI has just opened its most exciting chapter.
This article is from WeChat official account"Rongzhong Finance", author: Abu, editor: Wuren, published with authorization from 36Kr.