Momenta Goes Public, Auto Industry Team Building

On the day Momenta listed, the auto industry was unusually calm.

More than twenty automakers posted posters congratulating Momenta's listing on the same day. Including BMW, Mercedes-Benz, Audi, Cadillac, SAIC, GAC, etc., the scene was truly spectacular.

Beijing Benz CEO Rainer Dinkelmann specifically mentioned Momenta's listing at the Mercedes-Benz EQ GLC launch event, believing this is an important milestone defining the future of smart mobility. SAIC Volkswagen CEO Tao Hailong also congratulated Momenta's listing separately.

The auto industry's collective 'Team Building' is very rare.

First, in today's fiercely competitive Chinese auto industry, automakers are rivals to each other, so endorsing the same company is already rare.

The last time the auto circle had a 'Grand Unity' scene was during Mercedes-Benz's 140th anniversary, where many brands gathered to celebrate its birthday.

Second, multiple automakers posting posters to endorse the same supplier, celebrating its listing, is very novel and belongs to the first time in the Chinese auto circle.

Momenta receiving such favor from major mainstream automakers indicates that the relationship between Momenta and automakers differs from traditional client-vendor relationships. In the smart car era, Momenta has transformed from an automaker supplier to a 'infrastructure service provider'.

Among this, everyone reached a consensus: Major-brand intelligent driving will make cars sell better.

Momenta's ability to IPO is just outer information; the books, substance, and positioning that support Momenta's IPO are its industrial moat.

01. Momenta's Books, Three Special Features

Listing is a wonderful opportunity to peek at the core financial data of the first echelon of intelligent driving.

Opening Momenta's books, First, Momenta's financial data is very healthy, and closer to break-even than outsiders imagined.

From 2023 to 2025, Momenta maintained a triple-jump in revenue and gross profit. Company revenue grew from 743 million in 2023 to 2.413 billion in 2025, with an average annual revenue growth rate of over 80%.

At the same time, Momenta's gross profit margin climbed for three consecutive years, rising from 17.5% in 2023 to 71.6% in 2025.

One detail is worth noting, Momenta's revenue sources have changed significantly. Before 2024, Momenta was mainly focused on one-time project fee technical development services, where the number of projects undertaken in a year determined company revenue.

The turning point occurred in 2024, after which Momenta's main revenue became software licensing service revenue. The company can continuously benefit from automaker sales volumes. This means Momenta can rely on automakers to continuously repurchase and grow a snowball effect.

The change in revenue structure led to the improvement of the profit structure, this is the beginning of the business flywheel turning.

Therefore, in 2025, Momenta's adjusted net loss was only 303 million, close to the break-even point. Calculating at this pace, Momenta could achieve profitability first as early as this year.

Second, lies in the massive scale of this IPO.

This Momenta IPO raised a total of 5.89 billion HKD, ranking among the largest IPOs in the HK market, with a glamorous lineup of 14 cornerstone investors.

The Singaporean sovereign fund GIC among them deserves detailed explanation. GIC style is extremely conservative, rarely participating in new share cornerstone subscriptions, but this time GIC invested 100 million USD, becoming the largest share among cornerstone investors, its determination is evident.

So much so that Momenta was locked in advance before the IPO, the 14 cornerstone investors subscribed to 375.75 million USD in total, accounting for nearly 50% of the offered shares.

Third, the characteristic of customers being investors is extremely prominent.

Among the cornerstone investors, the names of Mercedes-Benz and BYD are particularly conspicuous.

These two automakers are both Momenta customers and investors. Data shows Mercedes-Benz contributed 25 million USD, BYD invested 15 million USD through associated entities.

Mercedes-Benz is the representative of the most demanding high-end manufacturing standards globally, BYD is the representative of the largest scale new energy mass production globally.

Two top automakers of different styles simultaneously chose the same supplier as a capital layout target, betting on the same target at the same time, meaning top automakers invested real money to identify the winner.

With the support of 40 million USD, there was a hint of reverse locking.

To this day, Momenta's intelligent driving is not a cost item, but a profit item. Automakers realize intelligent driving and even Physical AI will become core competitiveness, only by seizing the initiative first with capital to form a community of interests with suppliers can they seize the opportunity.

02. Only Data Flywheel Can Leverage Physical AI

After Momenta IPO, Cao Xudong's most memorable sentence was: 'I really believe in the data flywheel'.

Cao Xudong believes there are two kinds of belief: one is 'believing', the other is 'really believing'.

'Believing' means encountering problems without trying fully, retreating quickly to old methods,

'Really believing' means encountering all problems and first finding ways to change architecture and systems, making data-driven work.

The unchanging concept through the past decade of Momenta is the data flywheel.

The data flywheel can be broken down into three levels:

The first level is the 'One Flywheel, Two Legs' at the business level.

Using the data flywheel to feed back Momenta's two major business lines, one is mass production business, L2+ high-level assisted driving, L4 autonomous driving. The other is Scalable Robo, including Robovan, Robotruck and Robotaxi.

The reason why the data flywheel can drive two major businesses is that the technology concept of data-driven itself has universality.

Cao Xudong mentioned, Momenta has already verified on R7 World Model that the same set can support multiple downstream applications.

The second level is the 'Data Flywheel Acceleration' generated by data-driven.

From 2023 to 2026, Momenta technology experienced four stages: Rule Control, Deep Learning Control, Full Stack End-to-End, R6 Reinforcement Learning Model to R7 World Model.

The underlying logic is the same, data-driven models brought more reliable and predictable performance improvements.

This improvement has 'acceleration' characteristics. According to Cao Xudong's introduction, two years ago internal model performance saw 'tenfold improvement in two years', but the latest judgment is 'tenfold improvement per year'.

Expected by the end of the year, the R7 World Model will perform in par with the latest Tesla FSD.

The third level is the data flywheel leveraging Physical AI.

The data flywheel will also become the basic operation mode of Physical AI.

AI is divided into two categories, one is Digital AI, processing input and output as symbolic information, such as ChatGPT. The other is Physical AI, processing interactions with the real physical world, such as Momenta's World Model which predicts and reasons what will happen in the future by understanding laws and common sense of the physical world.

The reason Digital AI had the ChatGPT moment first is because internet data is easier to acquire.

Momenta listed as the first Physical AI stock, which is firm belief that Physical AI's GPT moment is coming soon.

The data flywheel is the food for the world model, using massive data from the physical world allows models to have common sense, which can then be reused to different business layers of Physical AI.

But as Jensen Huang said 'The physical world may be the larger part'.

The competitiveness of Physical AI compares who accumulates more and earlier real physical interaction data, whose model is more reliable. If the data flywheel is carried out to the end, the value ceiling of Physical AI is higher.

03. High-level Intelligent Driving Competition Window, Closing

Momenta's listing announced that the window period for high-level intelligent driving competition is closing, and the door to Physical AI is opening.

In the next 2-3 years, the mass production scale of leading autonomous driving manufacturers will grow exponentially, entering the stage where business revenue feeds back into robot R&D.

The core reason is that leading companies with first-mover and scale advantages have clarified the industry pattern, after 2026 the industry will trend towards convergence, the leading share will further expand.

From the current trend of Momenta, one can see a little.

Currently, mass production models equipped with Momenta assisted driving solutions have exceeded 1 million units, with cumulative targeted models exceeding 210 models.

According to CIC Zhongshi Consulting 'Autonomous Driving Industry Blue Book' shows, from March 2025 to February 2026, in the Chinese third-party urban NOA supplier market, Momenta's sales market share was 65%, among the top 10 automakers globally, 9 of them have already started cooperation.

One judgment is, scale itself will become obsolete, but the compounding speed of scale will not.

Momenta's highlights lie not only in scale, but also in the compounding speed of exponential growth.

From 2017 to 2024 Momenta lurked for 8 years, until the end of 2024 crossed the critical point and exploded with full force, that year it took Mercedes-Benz's all electric car and fuel car businesses at once.

2022 mass production 1 car model, 2023 about 8 car models, 2024 cumulative 20+ car models delivered. As of now, on-hand targeted models exceed 200. Calculating at this pace, expected to serve 10 million mass production cars by 2029.

The window left for mid-tier players to challenge the top status head-on is rapidly narrowing.

From a longer-term perspective, Momenta's listing this time previews the industry's upcoming moves - 2025 to 2027 is the peak period for high-level intelligent driving mass production on vehicles, after 2027 the industry will enter the stage where intelligent driving revenue feeds back into robot R&D.

Currently the robot industry is exceptionally lively, but mainly relies on Demo iterations and massive financing to advance.

Cao Xudong believes, limited by edge computing power and robot industry chain maturity, home robots will scale only by 2027, to do robots well, at least billions of USD investment is needed.

Momenta's strategy is very clear.

On one hand, improve base model capabilities, drive the performance of lower-layer applications of Physical AI.

On the other hand, as the Matthew effect in intelligent driving intensifies, leading intelligent driving companies can use revenue and gross profit to invest in robot R&D, the R&D model is healthier.

It can be foreseen, due to possessing technological universality, cash flow, what is truly worth expecting in the industry is actually automakers and intelligent driving players.