With the acceleration of intelligent driving towards L3 and above stages, the value center of China's automotive industry is undergoing structural changes.
In the past few years, the focus of capital market attention has been on algorithm models, chip computing power, and intelligent vehicle configuration. The core of industrial competition has begun to shift towards a more long-term value capability - security verification and large-scale testing capabilities.
In this trend, simulation verification platforms are no longer just research and development tools, but have become key infrastructure supporting the mass production and admission of advanced intelligent driving, and their strategic position has been greatly enhanced.
The "China Physical AI Simulation and Data Platform Research Report" released by global growth consulting firm Frost&Sullivan shows that the Chinese end-to-end advanced intelligent driving simulation and data platform market has entered a stage of concentrated acceleration. 51Sim, a subsidiary of Wuyi Vision, ranks first in the Chinese end-to-end advanced intelligent driving simulation and data platform market with a market share of 53.5%.
Autonomous driving: the core scenario for the first large-scale landing of physical AI
From a broader industry perspective, autonomous driving is becoming one of the most important and first large-scale application scenarios for "Physical AI".
According to Sullivan's prediction, with the accelerated expansion of downstream industries such as smart cars and embodied intelligent robots, China's physical AI simulation and data platforms are entering a significant period of value release, and their permeable market space is expected to reach over 180 billion yuan by 2030.
The so-called physical AI refers to intelligent systems that can perceive, make decisions, and act in the real physical world. Its core challenge is not the ability of a single algorithm, but how to continuously verify and optimize system security in complex, dynamic, and uncertain environments.
Autonomous driving naturally possesses typical physical AI features:
Should be oriented towards an open road environment;
Real time interaction with humans and transportation participants is required;
Must bear a high level of safety responsibility;
Therefore, its verification difficulty is much higher than that of pure software AI systems. In this context, simulation and data platforms have become the "training and validation factories" of the physical AI era. Only through high fidelity reconstruction, generative completion, and closed-loop simulation can extreme scenario testing on a scale of millions or billions be completed in a virtual environment, supporting algorithms to move towards safety and controllability.
This also explains why autonomous driving is considered the most important commercial breakthrough for physical AI - its scale, complexity, and regulatory requirements make it the most urgent area for simulation infrastructure needs.
Regarded as a capitalization representative of physical AI infrastructure
Wuyi Vision will be listed on the Hong Kong Stock Exchange in December 2025, with a net fundraising of HKD 650 million. The shareholders of Wuyi Vision include Moore's Thread, the first domestic GPU stock, SenseTime, an artificial intelligence enterprise, as well as Lightspeed, Cloud Nine Star VC、 Xie Nuo Investment and other funds. The May Day Vision has also attracted well-known investors such as Ge Weidong and Liang Botao to invest.
Wuyi Vision has formed a systematic layout in the field of physical AI simulation and data platforms, and is regarded by the market as one of the representative enterprises of the "first stock of physical AI".
Unlike single algorithm or single software product companies, Wuyi Vision's core capabilities focus on real-world digitization, physical environment reconstruction, and large-scale simulation verification system construction. This capability not only serves autonomous driving, but also has the potential to extend to more physical AI scenarios such as robots and embodied intelligence.
At the capital market level, the value logic of physical AI infrastructure enterprises differs significantly from that of traditional software companies
1， Stronger customer stickiness;
2， The cost of customer migration is higher;
3， Industry barriers are deeper;
Once entering the research and development system of the host factory and participating in the admission process, the platform often becomes a long-term partner rather than a phased tool procurement target. This' deep binding 'attribute enables platform based enterprises to have stronger long-term cash flow stability and potential for scale expansion.
Structural Concentration under the 'Strong Access' Cycle
With the continuous promotion of L3 access, car companies must provide a quantifiable, reproducible, and auditable safety verification system when releasing advanced intelligent driving functions.
Real road testing is limited by cost and risk, making it difficult to cover a large number of long tail scenarios, and end-to-end models are highly dependent on complex interactive environments. Simulation has become the only scalable verification method and a key technical support in the admission system.
Under this trend, the threshold for race tracks has significantly increased. The end-to-end simulation platform not only requires high fidelity physical and sensor modeling capabilities, but also massive data assetization capabilities, large-scale concurrent computing capabilities, and an engineering system deeply coupled with the R&D process of automotive companies.
Analysis suggests that the combination of high barriers to entry and strong market access has led to a rapid increase in industry concentration. 51Sim maintains its top position with a market share of 53.5%, surpassing the total of second to fourth places, demonstrating its head advantage. With the expansion of customer scale and the continuous accumulation of scene assets, platform capabilities are forming a network effect, and the industry landscape is accelerating stability.
The Long term Value Logic of Infrastructure based Enterprises
When advanced intelligent driving enters a new stage of parallel production and admission, industry competition has shifted from the competition of algorithm performance to the competition of system capabilities in data simulation verification systems.
Under this logic, platform based enterprises that can provide long-term support for autonomous driving access, safety certification, and continuous operational verification will become the underlying infrastructure providers in the era of intelligent vehicles.
Wuyi Vision and its subsidiary 51Sim represent not only a simulation software company, but also infrastructure builders in the era of physical AI.
As autonomous driving continues to expand as a core physical AI scenario, the strategic value of simulation and data platforms is expected to be further amplified.
For the capital market, what is truly worth paying attention to is no longer just the speed of algorithm iteration, but who can become the underlying platform supporting the long-term evolution of physical AI. From the perspective of market share, customer ecosystem, and engineering implementation capability, 51Sim is at the core of this industry trend, and Wuyi Vision is also establishing a clear top label on the physical AI infrastructure track.
