Physical AI Company Momenta Officially Listed, Market Cap Exceeds 70 Billion HKD

Momenta HK Listing: Why Does an Autonomous Driving Supplier Need to Talk About Physical AI?

Today, Momenta officially listed on the main board of the Hong Kong Stock Exchange, with an opening price of 301.000 HKD/share, 1.8% higher than the issue price. As of the mid-day break, Momenta's stock price was 305.800 HKD/share, with a market capitalization of 72.028 billion HKD.

Momenta HK Listing Scene, Image Source: Momenta

This is not the first time the autonomous driving industry has welcomed an IPO.

Before Momenta, Horizon Robotics had already completed its listing as an autonomous driving chip and computing solution company, while Pony.ai and WeRide also entered the secondary market sequentially with narratives as Robotaxi and L4 autonomous driving companies.

However, there is still a detail worth questioning regarding Momenta's listing this time: Why did a company whose main revenue still comes from passenger car smart driving solutions and whose core business still revolves around pre-installation mass production and high-level assisted driving redefine itself as a "global leading physical AI company" on the eve of its listing?

Momenta's prospectus shows that its revenue comes roughly from two parts: technology development services and software licensing. The former corresponds to customized smart driving solution development for automakers, while the latter corresponds to smart driving system usage fees after model mass production. In other words, Momenta's current most solid commercial foundation is still high-level assisted driving for passenger cars.

But in the new capital market narrative, what Momenta wants to tell is no longer just urban NOA, pre-installation mass production, and automaker customers, but a physical AI base model capable of understanding the physical world, deducing world evolution, and covering passenger cars, Robotaxi, Robovan, and even Robotruck.

Behind this is a collective identity shift occurring in the autonomous driving industry: When urban NOA changes from a technical highlight to a standard configuration for automakers, and when high-level smart driving solution providers fall into involution regarding price, customers, and mass production delivery, simply talking about "smart driving" is no longer enough to support higher imagination.

Thus, physical AI has become the new label, and also the new valuation language.

But questions arise: Is physical AI a natural extension of Momenta's technical route, or is it a new story that autonomous driving companies have to tell when the capital window narrows?

1. Behind the Listing

From the listing process perspective, the Hong Kong Stock Exchange was not Momenta's first choice.

In May 2024, Momenta secretly submitted listing application documents to the U.S. Securities and Exchange Commission (SEC), initiating a US IPO, planning to raise $200-300 million. At the time, the market and industry insiders expressed concern about its US listing, with emotions mainly concentrated on geopolitical and regulatory risks.

In 2024, during the period of escalating Sino-US tech competition, Chinese tech companies involving frontier technologies such as AI and autonomous driving faced increasingly strict scrutiny in the United States, and regulatory uncertainty rose significantly. Meanwhile, superimposed with the Federal Reserve's interest rate hike cycle leading to a global tech stock valuation correction, US market sentiment had already become cautious starting in the second half of 2024.

Although after Momenta submitted its listing application to the US market, the China Securities Regulatory Commission (CSRC) quickly issued a notice of record-filing for overseas issuance and listing, approving its US listing. However, since the validity period of the CSRC's overseas listing record-filing is 12 months, Momenta failed to complete the listing after it expired in June 2025, and the US listing plan was substantially shelved.

Momenta Overseas Issuance and Listing Record-Filing Notice, Image Source: CSRC

One point to note is that Chapter 18C of the HKEX Main Board Listing Rules had already taken effect in March 2023, allowing unprofitable tech companies to list on the HKEX main board, but Momenta still chose to aim for the US market in 2024 rather than immediately activating the promulgated Chapter 18C, indicating that between the US and Hong Kong markets, the former was still Momenta's first choice.

Why did Momenta spare no effort to promote the listing? From the outside, whether it is gross margin level, number of designated models, or adjusted net loss narrowing, it is better than other third-party smart driving system suppliers.

Prospectus content shows that Momenta's revenue is roughly divided into two parts: technology development services, i.e., customizing smart driving solutions compatible with automakers; and software licensing, i.e., charging smart driving system usage fees based on sales volume after model mass production. The former is related to the number of models cooperating with Momenta, while the latter is linked to model sales volume.

In 2025, Momenta's technology development service revenue was 1.45 billion RMB, and software licensing revenue was 970 million RMB. The company's adjusted net loss narrowed significantly from 1.093 billion RMB in 2023 to 303 million RMB in 2025, with the loss rate dropping from -147.2% to -12.6%, making it the closest to breakeven among similar companies.

Momenta Revenue Situation, Image Source: Momenta Prospectus

On the eve of the listing, on July 7, 2026, Momenta's official WeChat account posted that its smart driving system installation volume exceeded 1 million sets, further expanding its profit space for software licensing.

Momenta's revenue structure is also adjusting. In 2023, its technology service development revenue accounted for over 90%, but by 2025, software licensing revenue accounted for 40.1%. This is almost a "making money while lying down" business model, not only with marginal costs approaching zero but also capable of significantly raising Momenta's gross margin.

In 2025, Momenta's gross margin reached 71.6%. In comparison, Horizon Robotics' gross margin was 64.5% in the same year, Black Sesame Intelligence's was 41.0% and fluctuated significantly; WeRide's gross margin was 30.2%, and Pony.ai's was even lower, at only 15.7%.

Therefore, some industry views believe that Momenta does not lack money and does not seem to need to rush to list, but this assertion only holds true within the smart driving industry.

Although as of the end of 2025, Momenta's cash reserves on the books exceeded 10 billion RMB, for a listed company established for over ten years, its loss situation still cannot be taken lightly.

From 2023 to 2025, Momenta's cumulative net loss was as high as approximately 9.234 billion RMB. Although the adjusted operating loss is narrowing, the company expects to record a high net loss in 2026. From the operating logic of a listed enterprise, this situation cannot be called "not lacking money."

At the same time, as a technology-driven company, Momenta's R&D expenditure is huge. From 2023 to 2025, its R&D investment ratio was 172.4%, 113.8%, and 77.5% respectively. Although the ratio is decreasing year by year, the absolute amount of R&D investment is growing continuously, at 1.281 billion RMB, 1.508 billion RMB, and 1.869 billion RMB respectively from 2023 to 2025.

This means Momenta has hardly generated much substantive profit. Making money for R&D, and R&D for making money, this strange cycle is not what any company wants to see.

So, to achieve performance leap and maintain industry competitiveness, Momenta needs to tell a new story and also needs a new label.

2. Switching Identity

Momenta changed its label on the eve of the IPO: from "autonomous driving company" to "global leading physical AI company."

The label switch reflects changes in the industry competition landscape.

Momenta Official Website Corporate Introduction, Source: Momenta Official Website

From 2022 to present, China's smart driving industry has experienced a rapid development cycle from scale expansion to technical involution. In 2022, new force automakers such as XPeng, NIO, and Li Auto began pushing urban NOA functions first. In a market situation where "high-precision maps + rule algorithms" were still mainstream, what automakers and smart driving system suppliers focused on was the speed and scale of smart driving functions getting on board.

In the following two years, the maturity of the BEV+Transformer technical architecture allowed the industry to gradually eliminate high-precision maps and rule algorithms, making technical advantage the new battleground.

Entering 2025, the rise of technical paradigms such as end-to-end, VLA, and world models shifted the industry's focus back to who can first achieve "parking spot to parking spot" full-scenario intelligent driving, and who can make the car drive more like a human.

This speed of involution has brought a sense of urgency to all players, with no one able to guarantee they can survive smoothly. Momenta CEO Cao Xudong once publicly stated that the autonomous driving industry landscape will converge quickly, with only 2-3 companies in China and 3-4 globally ultimately winning.

If Cao Xudong's prediction is correct, Momenta's direct competitors will be players such as Huawei, Horizon Robotics, Zhuoyu, Yuanrong, etc., but if automaker self-developed smart driving systems are also counted, the competition Momenta faces is far more than just the contest between third-party suppliers.

In this situation, Momenta must quickly transform its identity, from a simple smart driving system supplier to a "technology pioneer that lets machines understand the world," even if this technology still cannot jump out of the scope of intelligent driving in the short term, or even cannot be applied to carriers other than passenger cars.

In April this year, Momenta mass-produced and launched the R7 World Model, a base model capable of understanding physical laws and deducing world evolution, capable of supporting the scaled landing of its products in fields such as passenger cars and Robotaxi.

Momenta Released R7 World Model During 2026 Beijing Auto Show, Image Source: Momenta WeChat Official Account

From the perspective of large models, the technical architecture of the R7 World Model includes model pre-training, model simulation, and reinforcement learning. Essentially, it is through massive data training to enable the model itself to form cognition and deduction capabilities of the physical world, giving it higher working efficiency in autonomous driving environments.

According to Momenta's plan, it can use one physical AI base model to simultaneously cover passenger cars, autonomous taxis (Robotaxi), and autonomous logistics vehicles (Robovan), and will expand to the autonomous truck (Robotruck) field in 2027.

But except for the passenger car field, Momenta's business in other segmented markets has hardly started. Although it has already obtained Robotaxi operation and testing permits in places such as Shanghai and Suzhou, it is still very far from scaled deployment.

In the prospectus, Momenta mentioned that "revenue from autonomous taxi service solutions is not material to Momenta's total revenue, because such solutions are still in the early stage of monetization."

Therefore, the blueprint of connecting multiple segmented markets with one physical AI model might, at this stage, be more aimed at making the capital market confident in it. This is not just Momenta's strategic goal, but something the entire industry is doing.

3. Staying at the Table

A noteworthy situation is that Chinese autonomous driving companies are almost all listing, and are all promoting the world model, a technical route that still has controversy. This is very rare in any industry.

Related research shows that as of the end of 2025, more than 15 companies in the autonomous driving field have updated their listing status. Pony.ai, WeRide, Cidi Intelligent Driving, Tudatong, etc., have successfully listed, while Yushi Technology, Mainline Technology, etc., are also sprinting to enter the secondary market. And after Momenta's listing, companies such as Qizhi Intelligent Driving, Yuanrong Qixing, etc., are also likely to land on the HKEX in concentration in the second half of 2026.

Why is there this situation of clustered listings? There are mainly two reasons: lack of money and involution.

From a time perspective, most Chinese smart driving companies were established between 2015 and 2018. For example, Horizon Robotics was established in 2015, Pony.ai, Momenta, and Black Sesame Intelligence were all founded in 2016, and WeRide was established in 2017. When these companies were established, they caught the wind of artificial intelligence and autonomous driving, and received a large amount of strategic investment from US dollar funds, well-known VCs, and OEMs.

But capital has cycles. According to the conventional duration of funds, RMB funds.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.