On October 22, at the 2026 World Intelligent Connected Vehicle Conference, Guo Shougang, Director of the First Equipment Industry Department of the Ministry of Industry and Information Technology (MIIT), stated that the second batch of market access pilot programs for Level 3 autonomous driving products is underway. The commercialization of Level 3 is actually a topic of high interest at this year's conference. Yang Dongsheng, Senior Vice President of BYD Group and Dean of the Automotive New Technology Research Institute, also expressed that Level 3 conditional autonomous driving is expected to be commercialized in 2028, while higher-level autonomous driving is expected to be commercialized after another one or two years of iteration based on this foundation.
As early as December 2025, two Level 3 autonomous driving models from Changan and BAIC had already received conditional market access approval. From the road pilot of the first batch of models to the promotion of the second batch of pilots, Level 3 is undoubtedly accumulating practical operational experience. However, the advancement of pilots does not mean that commercialization has been fully realized. There is still a distance between vehicles being approved for use under limited conditions and consumers being willing to pay for this capability, as well as automakers being able to continuously obtain commercial returns.
I am not pessimistic about the technical prospects of Level 3, but I remain reserved and skeptical about its short-term commercialization. The fact that the technology can operate does not equal consumers' willingness to pay; even if consumers are willing to use it, it does not mean automakers can recover their investments. Today, instead of discussing technology, I will simply explore the commercialization possibilities of Level 3 with everyone, and welcome you to share your views in the comment section.
01. Level 3 is Already on the Road, So Why Does This Not Indicate That Commercialization Has Been Fully Realized?
As mentioned earlier, on December 15, 2025, the MIIT conditionally approved two Level 3 autonomous driving models from Changan and BAIC. The Changan model can operate at a maximum speed of 50 km/h in traffic congestion scenarios on designated road sections in Chongqing; the Arcfox model can operate at a maximum speed of 80 km/h on designated highways and urban expressways in Beijing. It can be seen that these approvals have clear usage boundaries. The two models can only activate relevant functions on specified road sections and under corresponding operating conditions, and the designated operating entities will conduct road traffic pilots. It does not mean that Level 3 can be freely used on national roads after obtaining the approval.
On December 26 of the same year, the first batch of 46 Changan Deepal Level 3 pilot vehicles in Chongqing began pilot operations on designated road sections. This is significant progress in practical application, but pilot operations in specific areas cannot be equated with large-scale sales and use by ordinary consumers.
The fact that Level 3 can operate under specific conditions and the fact that Level 3 can serve a sufficient number of users are actually two different things. For consumers, the value of a driving function depends not only on whether it can operate, but also on how much of their daily travel falls within the system's operational design domain, and whether the function can be used conveniently and stably.
If a user's commuting route rarely passes through applicable road sections, then even if the vehicle is equipped with Level 3 capabilities, the actual benefits brought by this function may be limited. Consumers may not be willing to pay a significantly higher vehicle purchase price simply for an additional function with limited usage opportunities.
Of course, the current limited conditions do not mean that Level 3 can only operate on these road sections in the future. As product validation, market access, and related management mechanisms are steadily refined, the scope of application will definitely expand gradually. However, expanding the scope requires conducting corresponding validations for new scenarios and operating conditions, and also requires meeting corresponding safety and management requirements. Therefore, while the advancement of the second batch of pilots is indeed worth attention, it cannot directly prove that Level 3 is about to usher in large-scale commercial application.
What is truly worth attention in the future is how many real-world driving scenarios the product can cover, and whether this coverage can form stable user demand.
02. Will Consumers Be Willing to Pay If Automakers Invest More?
Will the investment from automakers inevitably drive consumers' willingness to pay?
The difference between Level 3 and Level 2 combined driving assistance is not just a higher degree of automation. When the Level 3 system is activated and within the operational design domain, the system is responsible for executing the dynamic driving task, and the driver no longer needs to continuously perform driving tasks as when using the Level 2 system. However, this does not mean the driver can ignore the vehicle under any circumstances. When the system issues a takeover request, the driver needs to respond as required; if the driver fails to respond in time, the system also needs to have corresponding risk mitigation capabilities. These requirements will affect the vehicle's system design, validation, and safety assurance. Automakers not only need to develop functions but also validate the system's safety performance within the operational design domain, handle system failures, driver takeovers, and risk disposal issues, and establish corresponding safety management mechanisms.
On July 30, 2026, the mandatory national standard GB 44721-2026 "Intelligent Connected Vehicles - Safety Requirements for Autonomous Driving Systems" was officially released and is planned to be implemented from July 1, 2027. This standard applies to Category M and Category N vehicles equipped with Level 3 and Level 4 autonomous driving systems, and specifies relevant technical requirements for autonomous driving systems, safety assurance requirements for vehicle manufacturers, and test conditions.
These safety requirements also indicate that automakers need to meet corresponding requirements in system R&D, validation, and operational assurance. However, the specific amount of increased costs depends on the enterprise's existing technical foundation, system solutions, and actual operational methods, and cannot be simply summarized with a fixed numerical value.
From a commercial perspective, can the R&D, hardware, validation, and safety assurance resources invested to meet these requirements ultimately obtain reasonable returns through higher selling prices, greater sales volume, or subsequent service revenue? If automakers include part of the costs in the vehicle's selling price, will consumers be willing to pay a higher price for Level 3? If it is not charged separately, but Level 3 is used as part of the model's competitiveness, can it help automakers increase sales and improve profits?
Currently, public information is not sufficient to prove that Level 3 has formed large-scale, stable paid demand, nor is it sufficient to judge how much incremental revenue automakers can obtain from it. More importantly, user demand for Level 3 will not automatically arise just because the technical level is higher.
For users who frequently drive on applicable road sections and can significantly feel the reduction in driving burden, this function may be more attractive; for users who rarely enter applicable scenarios, the benefits it brings may not be obvious. This means that Level 3 faces not only the issue of technical costs but also the issue of whether user value can cover the newly added costs. How much automakers are willing to invest in a technology cannot determine how much consumers are willing to pay for it. Even if system capabilities keep advancing, if the benefits perceived by users are not clear enough, technical advantages may not necessarily be translated into pricing advantages.
03. The Biggest Commercial Test for Level 3 is Whether Investment Can Be Transformed into Continuous Returns
For automakers, Level 3 certainly has strategic value. It can promote the R&D of higher-level autonomous driving, accumulate system safety validation experience, and may also drive the upgrade of the vehicle's electrical and electronic architecture, software capabilities, and safety systems. The value of these capabilities may not be immediately reflected in the profits of a specific vehicle model, but may be gradually realized in subsequent products.
However, strategic value and short-term commercial returns cannot be confused. Suppose an automaker invests more resources in developing Level 3, adds relevant hardware configurations, completes corresponding test validations, and obtains market access approval. If the function can only be used in limited scenarios and users' willingness to pay is unclear, there is uncertainty about whether this investment can be recovered through higher selling prices, greater sales volume, or subsequent service revenue.
Moreover, the safety requirements for Level 3 dictate that enterprises cannot only calculate the early-stage R&D investment. For Level 3, after the product is put into operation, the enterprise still needs to fulfill corresponding safety assurance obligations and accident liabilities, and must continuously handle problems exposed during operation.
As the scope of vehicle deployment expands, investments in operational monitoring, problem handling, and services may also increase; whether scale can amortize the per-vehicle cost depends on actual operational efficiency and usage scale. To expand the scope of Level 3 usage, automakers need to complete corresponding validations for new scenarios and operating conditions, and meet market access and safety management requirements; to cover more users, it is also necessary to deploy more qualified vehicles.
At the same time, enterprises must also prove that the newly added investment can be exchanged for sufficient user value and commercial returns. Scaling up may amortize the per-vehicle cost, but whether this can be achieved depends on the actual deployment scale, usage frequency, operational efficiency, and continuous assurance costs. If user demand is insufficient, merely expanding technical investment cannot naturally form better commercial returns. Therefore, one cannot judge that an automaker will definitely gain a significant commercial advantage in market competition simply because it is the first to obtain Level 3 market access.
What is truly worth discussing is which enterprises can transform technical capabilities in limited scenarios into products that users are willing to use, willing to pay for, and which can be continuously delivered.
04. Level 3 is Worth Looking Forward to, but Short-Term Commercialization Still Needs Market Proof
The second batch of Level 3 market access pilots is being promoted, indicating that the industry is still moving forward. As safety standards are gradually implemented and testing experience is steadily accumulated, the scope of future applications may expand, and product forms may also gradually enrich. However, before usage scenarios, users' willingness to pay, and enterprise profit models are more fully validated, one cannot judge that Level 3 is about to usher in a large-scale commercial explosion solely based on market access progress.
I am not pessimistic about the technical prospects of Level 3, but I do not think it will quickly become a good business. For automakers, obtaining market access is just the starting point.
The real test is whether, under the premise of meeting safety requirements, a sufficient number of users can continuously use this capability in a sufficient number of scenarios, thereby forming commercial returns that can cover long-term investments.
Being able to get on the road is technical progress; being able to continuously create commercial value is the true sign that commercialization is genuinely established.