Summary of Key Points
Momenta went public on the Hong Kong Stock Exchange as the "first stock in the field of physical AI," with its stock price closing flat on the first day at HK$295.6. Its market value of HK$69.6 billion fell short of the previously expected figure of HK$100 billion, indicating that the market has not yet assigned an additional premium to the "physical AI" label. The company, which was originally positioned in autonomous driving, suddenly shifted its narrative to focus on physical AI before the IPO. However, its main business strategy remains a dual-track approach: generating revenue from selling automotive software to support cutting-edge technologies like Robotaxi, while using production data to optimize algorithms. Nevertheless, the core capabilities of physical AI—such as cross-scenario migration and embodied intelligence—are still far from being realized, and its chip performance lags behind that of leading automakers. The current static P/E ratio of 26 times has already anticipated future prospects, and whether these expectations will be met depends on subsequent technological breakthroughs and revenue growth.
I. Signals from the First Day of Listing: Does the Market Buy into "Physical AI?"
Momenta's stock price soared but then declined on its first day of trading, ultimately settling at the issue price, resulting in a market value 30 billion Hong Kong dollars lower than the previously rumored figure of HK$100 billion. What does this indicate? Simply put, investors do not perceive the new identity as "the first stock in physical AI" to be worth more.
Why? Investors are focused on the company's actual business performance. Currently, Momenta's revenue mainly comes from automotive advanced driver assistance systems (ADAS) software, while Robotaxi is still in a small-scale pilot phase. Physical AI applications in areas such as home robots or healthcare services are virtually non-existent. The market is not naive; it will not pay more for a concept that has not yet been realized. Just because the company claims to be in the field of physical AI does not mean its valuation should be higher, especially since its core activities remain related to autonomous driving.
II. From Autonomous Driving to Physical AI: A Change in Narrative, Not in Business?
Momenta has always presented itself as an autonomous driving company until the week before its IPO when it changed its positioning to "physical AI." This is not a change in business model but rather a shift in the way it communicates its story.
Its core strategy remains a dual-pronged approach:
1. The Profit-Generating Leg: Selling ADAS software for mass-produced vehicles, with revenue expected to reach HK$2.4 billion by 2025 and a gross margin of 71.6%. The company has partnerships with major manufacturers like Mercedes-Benz and Toyota, which help fund its research and development efforts.
2. The Cost-Intensive Leg: Investing in the development of future businesses such as Robotaxi and Robovan, which are not yet generating revenue and may only become profitable by 2028.
The so-called "physical AI" actually involves extending autonomous driving technologies (such as the R7 world model) to various scenarios. For example, while there are plans for home robots, there is no product prototype available yet; these ideas remain at the CEO's level of discussion.
III. Is the "Hard Power" of Physical AI Enough?
To truly be a physical AI company, one needs tangible capabilities. Momenta's core technologies, the R7 world model and its self-developed BMC X7 chip, still need to be validated:
- R7 Model: Although claimed to be versatile, it is currently only used in automotive applications, and its ability to be applied to other scenarios (such as home robots) has not been tested. Moreover, since the training data is limited to road scenarios, it is uncertain whether it can be adapted for other contexts.
- BMC X7 Chip: With a performance of 272 TOPS, it is comparable to NVIDIA's previous-generation Orin X chip but falls short of leading domestic automakers' chips (e.g., Xpeng's Turing 750TOPS and NIO's Shenji 1000TOPS). As future large models require more computing power, this chip may not be sufficient.
In essence, the current technology is only suitable for autonomous driving and is far from meeting the requirements of true physical AI across different scenarios.
IV. Valuation Bubble: Can It Be Realized in the Future?
Momenta's current static P/E ratio of 26 times exceeds even Tesla's highest valuation during its peak in 2021. This suggests that the market is pricing it as a future physical AI giant, despite its actual role being that of an autonomous driving technology provider.
The question is whether it can sustain this valuation:
- Short Term: Revenue from mass-produced ADAS may continue to grow, but at a slower pace due to the limited size of the automotive market.
- Long Term: Success will depend on the commercialization of Robotaxi and physical AI applications. While Cao Xudong predicts a market explosion for Robotaxi by 2028, there are only dozens of vehicles in trial operation; scaling up to 100,000 units by that time is challenging. Home robots are even further off, with potential mass adoption not expected until 2030.
If revenue growth fails to keep up with the valuation, Momenta's market value could shrink significantly. After all, investors will not continue to support a company based on unproven promises.
V. Advantages and Limitations of Being an Independent Third-Party Supplier
As an independent third-party provider, Momenta has an advantage: automakers that wish to avoid dependence on Huawei or reduce risk by not developing their own technologies may choose it. Partnerships with companies like Mercedes-Benz and SAIC contribute to the stability of its mass production business.
However, there are also limitations. Without its own car brand, Momenta relies on automakers for revenue. If other manufacturers develop their own autonomous driving systems (e.g., Xpeng or Li Auto) or partner with other suppliers (e.g., Horizon Robotics), Momenta's market share could be threatened. Additionally, physical AI requires data from various scenarios, and relying solely on automotive data may not be sufficient to expand into home or industrial applications.
In summary, Momenta's status as the "first stock in physical AI" is more of a narrative upgrade for the IPO. To truly become a physical AI company, it must demonstrate its ability to apply automotive technology to other areas and address challenges related to chip performance and cross-scenario capabilities. The current valuation already reflects high expectations, and the future will determine whether these promises can be fulfilled.