Summary of Key Points
Over the past six months, three companies from different sectors (Wuyi Vision, Haiqing Zhiyuan, and Momenta) have all claimed to be the "first stock in physical AI." Behind this is capital seeking new stories amid the intense competition in generative large models, with physical AI becoming a hot topic. However, these companies have vastly different core businesses (digital twins, multi-spectral perception, autonomous driving), and they can all fit under the broad concept of physical AI due to its ambiguity. Although the "physical AI" label can bring short-term valuation premiums (for example, Haiqing Zhiyuan's stock price rose 270% on its first day of trading), the market will return to focusing on performance, and such labels will not maintain their premium status in the long run.
Why Do They All Claim to Be the "First Stock in Physical AI?"
The competition in generative large models (such as ChatGPT) has been too fierce in the past two years, with top companies attracting most of the funding. Smaller firms have struggled to raise capital, and even the leading ones are losing money (Momenta lost 9.2 billion yuan in three years). Capital is in need of new narratives, and physical AI has been labeled by NVIDIA's Jensen Huang as the "next wave of AI," with a market size estimated at $6 trillion, making it a new target for investment.
More importantly, the term "physical AI" is more appealing than previous categories. Calling oneself a "digital twin company" may go unnoticed, and an "autonomous driving solution provider" might not attract much interest, but adding the "physical AI" label can significantly boost valuation. It's like selling a product; a simple package might sell for 10 yuan, but with a "hot trend" label, it could fetch 50 yuan—everyone wants to use this label to raise more money and avoid a stock price drop at launch.
Three Companies with Different Focuses, Yet All Claiming Physical AI
The definition of physical AI is broad: it combines physical laws with AI, either simulating real-world phenomena in virtual environments or enabling machines to perceive and understand the physical world. This flexibility allows various companies to fit into the category:
- Wuyi Vision: They are working on a "Earth clone project," which involves replicating the Earth in a virtual world (digital twin), thus fitting under physical AI.
- Haiqing Zhiyuan: They use ultraviolet and infrared multi-spectral signals to enable machines to "see" the physical world, which also falls within the scope of physical AI.
- Momenta: Their autonomous driving algorithms understand physical properties like mass, speed, and friction, which is again considered part of physical AI.
The ambiguity of the concept means anyone can claim to be the first—just by adding appropriate qualifiers. For instance, Wuyi Vision adds "global," Haiqing Zhiyuan adds "multi-spectral," and Momenta adds "foundation model," turning the title of "first stock" into a rhetorical competition.
Has the Title of "First Stock" Become a Short-Term Trend?
The concept of "first stock" has shifted from being about who launches first to a game of adding descriptive qualifiers:
- Wuyi Vision plans to go public by the end of 2025 and claim to be the "world's first physical AI stock."
- Haiqing Zhiyuan went public in June 2026, omitting "global" and directly calling itself the "first physical AI stock."
- Huijie Electronics took it a step further by adding "edge-side wireless intelligent chips," becoming the "first physical AI edge-side wireless intelligent chip stock."
Timing is more crucial than being the first to launch. When Wuyi Vision went public, physical AI was not yet a hot topic, and only institutional investors were interested; Haiqing Zhiyuan emerged during a period of market excitement, with retail investors buying 7,181 times their shares (equivalent to HK$440 billion), causing its stock price to rise 270% on the first day, making it seem more legitimate as the "true first."
Can Labels Generate Value, but Only for a Short Time?
Labels can significantly boost valuation in the short term. For example, Haiqing Zhiyuan's stock price surged due to the "first physical AI stock" label, but it dropped 26% on the third day of trading, reflecting that investors realized the label was for speculation. As the market matures, performance becomes the key factor.
The Anxiety Behind the Companies
These companies are eager to go public because they can no longer raise enough funds in the primary market (private equity). In 2025, funding for large models decreased by 23% in both volume and value, with only the top players getting significant investments. Early investors' funds are about to expire, forcing them to exit through IPOs. The Hong Kong Stock Exchange's 18C rule allows unprofitable tech companies to list, providing a last resort.
The 18C rule has made Hong Kong a crucial option for these companies. However, while going public is easy, selling shares is challenging if they rely on traditional themes (autonomous driving, digital twins). The "physical AI" label helps increase valuation and attract retail investors, ensuring sufficient funding.
Will the Label Game Continue?
The trend of using labels to gain a competitive edge will likely continue for a while, but the window of opportunity is limited. Once the label loses its appeal, the market will focus on the companies' actual capabilities. For investors, it's important to assess the company's real offerings rather than being misled by short-term hype. For companies, labels are just a stepping stone; ultimately, their technology and profitability determine their success.
In conclusion, while labels can open doors to the capital market, they cannot sustain long-term value. Companies must prove their worth through tangible results and technology. The race for the "first stock" title is more about timing and marketing than actual innovation.