Autonomous driving has reached the "Hong Kong stock moment": What signals are being sent behind Pony.ai's second listing?
Entering 2025, global capital began to vote for autonomous driving with real money. Subsequently, the industry has moved from the technology verification stage to the critical point of large-scale commercialization, and the signals of industrial transformation have become increasingly clear. The significance behind Pony.ai's listing on the Hong Kong Stock Exchange lies not only in raising development funds for the company itself but also in providing an observational sample and a cornerstone of confidence for the entire autonomous driving industry. At this critical inflection point where technological transformation intersects with the capital cycle, the revaluation of the value of the autonomous driving industry has just begun.
The autonomous driving track has announced the arrival of the "Hong Kong Stock Moment".
On November 6, Pony.ai, a leading domestic autonomous driving company, successfully listed on the Hong Kong Stock Exchange, setting the largest IPO scale for global autonomous driving in 2025 and becoming the largest IPO in the AI field of the Hong Kong Stock Exchange in 2025.
It has been less than a year since it rang the bell on the NASDAQ stock market. From being the "world's first Robotaxi stock" on the US stock market to the "first autonomous driving stock on the Hong Kong Stock Exchange", Pony.ai has completed a rapid and in - depth leap in the capital market, marking a key step in its "US stocks + Hong Kong stocks" global dual - home strategy.
Notably, three months before Pony.ai's listing on the Hong Kong Stock Exchange, the ARKQ fund under Cathie Wood, known as the "female Warren Buffett", first invested in a Chinese L4 - level autonomous driving company. In just about a month, ARKQ increased its stake in Pony.ai (NASDAQ: PONY) three times, making the capital market smell a familiar scent - it is very similar to Cathie Wood's investment style in the early stage of Tesla.
Why are international capitals re - betting on autonomous driving at this time? Does Pony.ai's secondary listing indicate the arrival of an industry inflection point?
"Cathie Wood" Increases Stake Multiple Times, and Growth - Oriented Technology Stocks Usher in Value Revaluation
A detail is that the ARKQ fund under Cathie Wood bought shares on the day after Pony.ai released its second - quarter financial report. It spent $12.9 million on the initial position. Less than a month later, on September 17 and 18, it added 88,335 shares and 31,764 shares respectively.
From the performance report, Pony.ai showed a strong momentum in the first half of this year. The financial report shows that in the first half of 2025, the company achieved a revenue of $35.43 million (approximately RMB 254 million), a year - on - year increase of 43.3%. The Robotaxi segment, as the core business, was particularly eye - catching. The revenue in the first half of the year reached $3.256 million (approximately RMB 23.32 million), a significant year - on - year increase of 178.8%.
Cathie Wood is not the only one attracted by the outstanding performance of this "dark horse". Since the second quarter, Baillie Gifford, a British capital, has invested more than $100 million to take a large stake. Moreover, 14 international top - tier institutions, including Fidelity Investments and Wellington Management, have collectively increased their stakes. Pony.ai has become the "core target" for global capital to bet on the autonomous driving industry.
Among the international capital giants, Cathie Wood and Baillie Gifford have attracted the most market attention. The former is well - known for capturing "disruptive innovation", and ARK Invest under her focuses on investing in technology stocks with "disruptive innovation potential". The latter is a century - old asset management giant that has witnessed the birth of numerous industrial growth - oriented technology stocks.
Both are iconic institutions known for long - term investment. In history, they jointly made large - scale bets on Tesla and NVIDIA in their early stages and then obtained returns of dozens of times. Therefore, they are once regarded as the "wind vane" for observing the development potential of global growth - oriented technology stocks.
In terms of model, the stage characteristics of Pony.ai are highly similar to those of Tesla from 2009 to 2012: continuous investment during the technology climbing period, financial pressure, but rapid growth in R & D intensity and patent reserves. This "early non - consensus" structural opportunity has led to the coincidental choices of the two "wind vanes" again.
On the other hand, in recent years, the Hong Kong stock market has provided fertile ground for the development of Chinese technology companies' globalization strategies in terms of policy support, system optimization, valuation advantages, and international resources.
In the current macro - interest rate environment and the context of the AI narrative, the valuation logic of growth - oriented technology stocks is also undergoing profound changes. The low - interest rate environment provides more favorable financing conditions for long - cycle and high - investment technology companies, while the breakthrough progress of AI technology has injected new impetus into frontier fields such as autonomous driving.
Therefore, international funds have begun to re - evaluate the long - term growth potential of the autonomous driving sector. Behind the concentrated layout of capital is not only the capital recognition of Pony.ai as a benchmark enterprise but also the anticipation of the upcoming explosion of commercialization in the Chinese Robotaxi industry.
In the latest report "ARK Invest Big Ideas 2025" released by the ARK fund, Robotaxi is regarded as one of the most disruptive industries in the next five years. It is estimated that by 2030, the global Robotaxi market scale will reach $10 trillion, and the total industry valuation will be as high as $34 trillion.
This judgment of the development prospect has been echoed by many international institutions. In recent years, the autonomous driving industry has always attracted the attention of capital and has also experienced bubbles and troughs. Cathie Wood once pointed out that "the key to the success or failure of Robotaxi lies in whether the cost model can work". Now, his various actions have shown that the time is almost ripe.
Accelerated Dual - Line Commercialization, Cost Breakthrough Drives the "Profit Timeline"
In fact, the success or failure standard of whether the cost model can work is applicable to most players in the technology track. From the development law of growth - oriented technology companies, R & D investment, periodic losses, and low - gross - margin normalcy are the inevitable stages before commercialization.
To make the cost model work, technology companies must first outperform in terms of funds, technology, and the market. Therefore, the mass - production road - testing ability has become one of the important dimensions for measuring the strength of Robotaxi enterprises.
Currently, Pony.ai has more than 720 Robotaxis and more than 170 Robotrucks. It is expected that the number of Robotaxis will reach a scale of one thousand by the end of the year. Peng Jun, the CEO of Pony.ai, believes that "autonomous driving must be developed in China, which has the world's richest and most complex road data. Testing can feed back into R & D."
In China, Pony.ai is currently the only enterprise that has obtained permits for driverless travel services and carried out fully driverless paid operations in the four first - tier cities of Beijing, Shanghai, Guangzhou, and Shenzhen. At the same time, its Robotaxi services in Guangzhou, Shenzhen and other places have been operating 24/7.
As of 2025, Pony.ai has accumulated more than 70 million kilometers of testing and operation mileage, including more than 10 million kilometers of fully driverless road - testing mileage.
If the mass - production road - testing ability is the foundation for the large - scale commercialization of Robotaxi, then technological drive is the most critical engine for making the cost model work.
Since June, the seventh - generation L4 - level Robotaxis jointly developed by Pony.ai with BAIC, Toyota, and GAC have started mass production. They have now completed the production capacity ramp - up and entered stable production, and have obtained L4 - level autonomous driving demonstration permits in Beijing, Shanghai, and Guangzhou.
As an L4 - level solution with the mission of large - scale and commercialization of fully driverless driving, the seventh - generation Robotaxi model no longer uses industrial control computers but is equipped with Pony.ai's self - developed vehicle - grade domain controller, which is composed of four NVIDIA OrinX chips. From wiring harnesses, connectors to solid - state lidars, chips, etc., 100% of the components are vehicle - grade.
Pony.ai is the world's first and currently the only company to use vehicle - grade SoC computing chips to achieve L4 - level fully driverless Robotaxis. The full vehicle - grade standard has rapidly reduced the bill of materials (BOM) cost of the seventh - generation autonomous driving hardware kit, a significant 70% reduction compared to the previous generation. The design life of the entire hardware system is up to 600,000 kilometers over 10 years, and most of the seventh - generation models have been pre - installed and mass - produced.
A research report by Citibank in September 2025 shows that among the global Robotaxi enterprises that have achieved fully driverless external operations, the BOM cost of Pony.ai's seventh - generation Robotaxi autonomous driving system solution is the lowest.
Zhang Ning, the vice - president of Pony.ai and the person - in - charge of the Robotaxi autonomous driving travel business, said that "in cities like Beijing, Shanghai, Guangzhou, and Shenzhen, when the number of Robotaxis reaches 1000, the operation will reach the break - even point. Beyond this point, for each additional vehicle, our cost will be lower, the gross profit margin will be higher, and we will enter a positive self - financing stage."
In the Robotruck business, Pony.ai sees a clearer profit path:
For B - end business scenarios such as trunk lines and dedicated lines, it re - uses 80% of the Robotaxi technology to launch autonomous driving freight services, achieves single - line profitability in the B - end first, and then feeds back and supports the long - term technological iteration and market expansion of Robotaxi. Currently, the company's autonomous driving truck freight volume exceeds 1.07 billion ton - kilometers, and the autonomous driving mileage is approximately 6.8 million kilometers.
With the accelerated dual - line commercialization, Pony.ai has successfully crossed the "technology verification" stage and entered the acceleration cycle of large - scale operation, and the "profit timeline" has become clearer.
Financial data confirms this trend. As its core business, the passenger fare revenue of Robotaxi soared by approximately 800% and more than 300% year - on - year in the first and second quarters of 2025 respectively, and the average daily order volume of fully driverless Robotaxis exceeded 15.
According to the current commercialization speed, Pony.ai expects to achieve break - even for single - vehicle operation in 2025 and start the journey towards large - scale commercialization of Robotaxi.
"World Model" Builds the Core Moat, and Technological Barriers Are Transformed into Double Dividends
In fact, while the capital market is cheering for Pony.ai's commercialization progress, its technological core, the "World Model (PonyWorld)", is quietly building an insurmountable moat.
The World Model is a simulation closed - loop of the physical world. This simulation training environment generated based on the reinforcement learning paradigm is not a single model but a complete autonomous driving system covering perception, decision - making, and control. It is more like an "AI driving school" tailored for in - vehicle models.
Pony.ai has created a "virtual driver" suitable for various vehicle models and application scenarios based on this. With the three core capabilities of the World Model, namely generating scenarios and sensor data, highly realistic simulation, and multi - dimensional evaluation of driving behavior, as well as the support of tens of billions of kilometers of test data generated every week, the virtual driver can achieve self - evolution under multiple paths.
This powerful technological barrier has also been directly transformed into double dividends of safety and cost.
In terms of safety, Pony.ai's autonomous driving system has achieved zero accidents in 500,000 hours of full - scenario, all - weather, and fully driverless operation. The takeover rate is as low as 1 time per million kilometers, far better than the industry average of 1 time per 100,000 kilometers. The safety performance of L4 autonomous driving is 10 times higher than that of human driving.
Behind the amazing operation data, the L4 - level safety redundancy system has been comprehensively upgraded. After being fully verified in the simulation environment and real scenarios, the newly upgraded safety system can effectively handle various single - point failure scenarios and provides a safety foundation for driving safety.
In terms of cost, the virtual training provided by the World Model has significantly reduced the investment in vehicles and manpower for real road - testing and also greatly shortened the R & D cycle of the new - generation autonomous driving system. It only took 7 months from the global debut of the seventh - generation Robotaxi in April this year to its official operation in November of the same year. While compressing the R & D cost, it also demonstrates the rapid iteration ability of L4 - level autonomous driving technology.
This competitiveness of reducing R & D costs with technological support is also reflected in Pony.ai's self - developed hardware capabilities. Its self - developed domain controller based on four NVIDIA Drive Orin - X chips has a computing power of 1016 TOPS. Compared with the previous generation, its volume, weight, and power consumption have been reduced by more than 50%, and the cost has been reduced by 80%. It has become the world's first L4 full - scenario mass - production domain control solution that can be commercially available on a large scale.
It is worth mentioning that Tesla, known for its disruptive innovation, recently disclosed a new project called the "World Simulator". This system aims to create an infinitely realistic virtual training ground for its autonomous driving and robot projects, which has something in common with the "World Model".
Pony.ai completed the technological route transformation as early as 2020, adopting the "learning by practicing" (virtual training) model, constructing this World Model capable of generating infinite scenarios, and starting iterative evolution. It is this sense of independent innovation that has built the foundation of Pony.ai's technological moat.
From Pony.ai's capital operation of dual - listing on the Hong Kong Stock Exchange, to the accelerated promotion of dual - line commercialization of Robotaxi and Robotruck, and then to the technological moat built by the World Model, the development trajectory of this leading autonomous driving enterprise clearly marks the evolution path of the industry - autonomous driving is moving from the technology verification stage to the critical point of large - scale commercialization.
[Conclusion]
The intensive layout of international capital, the continuous verification of the business model, and the continuous strengthening of technological barriers all point to one conclusion: the "Hong Kong Stock Moment" of autonomous driving is not only a capital milestone for an enterprise but also an important sign that the entire industry is ushering in value revaluation.
Through the dual - listing structure of "US stocks + Hong Kong stocks", Pony.ai has not only won the recognition of international capital but also opened the participation channel for mainland funds, building a solid financial foundation for its subsequent development. More importantly, its practice shows that the commercialization of autonomous driving is no longer an unattainable vision.
This article is from the WeChat official account "Digital Science Society" (ID: sktxs0), author: Beiye. It is published by 36Kr with permission.