On October 29, Zhuoyu Technology officially launched its commercial vehicle technology brand "Zhuoyu Panxing" and announced collaboration progress with seven major enterprises: XCMG Auto, Shaanxi Automobile Group, China National Heavy Duty Truck Group (Sinotruk), FAW Jiefang, Foton Motor, Dongfeng Commercial Vehicle, and Yutong Bus. The business scope covers heavy-duty trucks, buses, light-duty trucks, and other vehicle types.
On the same day, the first batch of mass-produced L4 autonomous driving heavy-duty trucks, the T9 ROBOTRUCK jointly developed by GAC Lingcheng and Pony.ai, officially rolled off the production line. Looking back further, companies such as QCraft, Minieye, and Desay SV Automotive have also successively entered the fields of unmanned logistics and low-speed unmanned delivery.
In the past few years, the primary competitive goal for ADAS suppliers was to enter the supply chains of more passenger vehicle brands and secure projects for urban NOA, highway NOA, and integrated driving-parking systems. Now, an increasing number of companies are extending their technology and business into areas such as heavy-duty trucks, light-duty trucks, buses, and logistics vehicles.
This does not mean that the passenger vehicle ADAS market has lost its value, but rather that suppliers are facing an issue they cannot ignore.
As technical barriers in passenger vehicle ADAS continue to rise, prices remain under pressure, and automakers strengthen their self-research capabilities, is concentrating all resources on the passenger vehicle market still the best choice?
01. Passenger Vehicle Market Is Still Growing, But It Is Becoming Harder for Suppliers
From a market size perspective, it is clearly premature to say there are no opportunities left in passenger vehicle ADAS.
Advanced driver-assistance systems are still being popularized across more price segments, and urban NOA is gradually trickling down from high-end configurations to mid-range models.
For consumers, ADAS features are becoming increasingly important factors when purchasing a car. However, for suppliers, market growth and their own profitability are not the same thing.
In the past, passenger vehicle ADAS suppliers could establish competitive advantages through technological capabilities. Whoever achieved highway NOA earlier, implemented urban NOA first, or delivered better system performance in complex road conditions had the opportunity to win design-in orders from automakers.
Now, the competitive landscape has changed.
On one hand, more and more automakers are establishing their own ADAS R&D systems. Some choose full-stack self-development, while others adopt a hybrid approach combining self-research and supplier collaboration. For suppliers, even with mature algorithms and mass-production experience, they must continuously prove that they can provide technical value that automakers find difficult to achieve independently or unnecessary to duplicate.
On the other hand, as ADAS features accelerate in adoption, competition among suppliers has shifted from merely comparing functions to a comprehensive contest of performance, cost, mass production capability, and adaptation efficiency.
In the past, a set of advanced ADAS solutions could enter high-end models based on functional advantages; now, automakers not only care about whether the system can achieve urban navigation, but also compare chip costs, sensor configurations, computing platforms, development cycles, and subsequent maintenance costs.
This places higher demands on suppliers.
Especially as urban NOA gradually descends to lower-priced models, suppliers need to continuously improve system capabilities within limited hardware costs and price margins.
A drop in solution prices does not mean that R&D costs can decrease proportionally.
Model adaptation, data collection, model training, road testing, and mass-production verification still require continuous investment. Although platform reuse can reduce some marginal costs, it is difficult to completely eliminate these investments.
For third-party ADAS companies, this pressure comes not only from competitors but also from customers themselves.
When automakers possess stronger self-research capabilities, suppliers no longer naturally hold a technological advantage; when ADAS features become standard equipment in more models, suppliers struggle to gain long-term premiums based solely on specific functions.
This is why, although the passenger vehicle ADAS market is still growing, suppliers are beginning to seek other business opportunities.
What they need is not just more design-ins, but new revenue sources and more sustainable business models.
02. The Appeal of Commercial Vehicles Goes Beyond Just Another Market Segment
Compared to passenger vehicles, commercial vehicle ADAS has a distinct characteristic: its technological value is easier to link with operational benefits.
Passenger vehicle consumers buying ADAS features focus on driving experience, safety, comfort, and convenience. How much driving burden the ADAS system reduces or how smoothly the vehicle drives is hard to directly convert into clear economic gains.
Commercial vehicles are different.
For a heavy-duty truck operating long-term, fuel consumption, transportation efficiency, driver workload, vehicle safety, and operating costs are issues that fleet operators face every day.
If an assistance driving system can comprehensively consider road gradient, curvature, vehicle load, and powertrain status to reasonably plan speed, gear shifting, and braking strategies, it has the potential to reduce unnecessary acceleration/deceleration and energy loss; if it can alleviate the operational burden on drivers during long-distance transport, it may also improve fatigue management and transportation efficiency.
These values feed directly into the cost accounting of logistics companies.
Zhuoyu's recent concentrated launch of the "Panxing" commercial vehicle technology brand is a typical case. Its PDS 4 primarily targets heavy-duty truck trunk logistics scenarios and supports highway NOA. Zhuoyu disclosed that the Shaanxi Auto Delong X5000 equipped with this solution went into mass production in June 2026, with hundreds of units currently deployed in express delivery trunk operations.
Regarding fuel-saving effects, test data published by Zhuoyu shows that its AI reinforcement learning fuel-saving algorithm achieved a measured fuel saving rate of 4.5%–7.5% in tracking statistics covering 230 mass-produced vehicles and over 20 million kilometers cumulatively. According to corporate estimates, each vehicle can save approximately CNY 20,000 to CNY 30,000 in fuel costs annually.
It should be noted that the fuel saving rate and annual benefits are test and estimation results disclosed by the company. Actual effects are influenced by routes, loads, driving habits, traffic conditions, and vehicle working conditions, so it cannot be inferred that all heavy-duty trucks will achieve the same level.
However, this set of data at least indicates that the value of commercial vehicle ADAS does not necessarily rely on consumers' subjective feelings about intelligent features but can also be measured by operating costs.
For suppliers, this method of expressing value is important.
If an ADAS system can help logistics companies reduce operating costs, customers can evaluate whether the investment is worthwhile by combining vehicle procurement costs, system prices, mileage traveled, and actual savings.
In comparison, while the value of passenger vehicle ADAS is equally important, how much premium consumers are willing to pay for a feature, and how much automakers are willing to pay for supplier solutions, are heavily influenced by brand positioning, whole-vehicle pricing, and market competition.
Another attraction of commercial vehicles is that certain application scenarios are relatively concentrated.
The operating routes and transport tasks for heavy-duty truck trunk transportation are relatively clear. Some scenarios allow for a more defined Operational Design Domain (ODD), facilitating suppliers in developing and verifying functions around limited routes, speeds, and road conditions.
But this does not mean heavy-duty truck ADAS is simpler.
Heavy-duty trucks have large masses, long braking distances, and significant load variations. Their vehicle dynamic characteristics differ greatly from passenger vehicles. Scenarios such as long downhill slopes, continuous curves, ramp merging, and temporary construction still place high demands on the system's perception, decision-making, and control capabilities.
From this perspective, commercial vehicles are not substitutes for passenger vehicle ADAS, but rather represent a different logic of value.
03. Suppliers Are Not Targeting the Same Type of Commercial Vehicle
If we look at the recent actions of ADAS companies together, we find that while they are all expanding beyond the passenger vehicle market, the directions they choose are not entirely identical.
Zhuoyu's recent concentrated release of its commercial vehicle technology brand and cooperation progress focuses on extending the technical capabilities accumulated in the passenger vehicle domain to heavy-duty trucks, buses, and light-duty trucks. Their solutions do not simply transplant passenger vehicle algorithms onto heavy-duty trucks; instead, they use existing model and data foundations as a starting point, combining heavy-duty truck-specific data, scenario enhancement, and vehicle dynamics adaptation to develop assistance driving systems for trunk transportation.
Zhuoyu disclosed that its accumulation of passenger vehicle driving data exceeds 3 billion kilometers, while it possesses 400 million frames of heavy-duty truck-specific data. The list of partners announced covers major domestic commercial vehicle manufacturers, indicating its desire to leverage commercial vehicle OEMs' model platforms and mass-production capabilities to further establish a foundation for scaled delivery.
But Zhuoyu is not the only company doing this.
On October 8, 2026, Foton Motor and Zhuoyu Technology signed a strategic cooperation agreement. Both parties will collaborate on advanced intelligent driving for commercial vehicles, covering directions such as highway NOA for heavy-duty trucks, national road assisted driving, intercity highway NOA, and urban NOA for light commercial vehicles.
Pony.ai has chosen a different path.
On October 29, the first batch of mass-produced L4 autonomous driving heavy-duty trucks, the T9 ROBOTRUCK jointly developed by GAC Lingcheng and Pony.ai, rolled off the line. Unlike Zhuoyu, which mainly focuses on highway assistance driving for heavy-duty trucks, Pony.ai's layout emphasizes L4 autonomous freight transport, with clear differences in technical goals and operational models.
The value of this route lies in the fact that autonomous driving technology is no longer just helping drivers complete partial driving tasks, but attempting to let vehicles undertake more transport tasks under limited conditions.
However, it needs to solve not only the vehicle's autonomous driving capability but also issues related to operational organization, fleet scheduling, remote assistance, exception handling, and transport networks.
The rollout of the first batch of mass-produced models is merely a step from R&D manufacturing to product verification and scenario application, and does not indicate that L4 heavy-duty trucks have completed large-scale commercial verification.
Unmanned logistics vehicles represent another route worth noting.
QCraft reached a strategic cooperation with Chery Commercial Vehicle on unmanned logistics as early as November 2025. The two parties plan to create mass-production-grade L4 logistics products.
The first unmanned logistics vehicle from this cooperation has been gradually put into operation in cities such as Suzhou, Jinhua, and Wuhu, attempting to combine autonomous driving technology with commercial vehicle mass-production manufacturing and urban delivery demands.
Minieye signed a strategic cooperation agreement with Didi Freight Delivery on full-domain unmanned logistics operations in July 2026, providing Xiaozhu series L4 unmanned logistics vehicles, autonomous driving software systems, and technical support.
Both parties hope to combine freight platform resources with unmanned vehicle mass-production capabilities to promote the scaled implementation of last-mile delivery. Public information shows that Xiaozhu series unmanned vehicles have been deployed in multiple cities.
Desay SV Automotive also launched the low-speed unmanned vehicle brand "Chuanxing Zhiyuan" in September 2025, introducing the S6 series low-speed unmanned vehicles, beginning to extend its automotive electronics and intelligent driving technology to low-speed unmanned vehicle scenarios such as parks and logistics distribution.
The layouts of these companies indicate that ADAS technology is extending from passenger vehicle assistance driving to more mobile scenarios.
However, it must be clarified that heavy-duty truck highway NOA, L4 autonomous heavy-duty trucks, urban unmanned logistics vehicles, and low-speed delivery vehicles have obvious differences in technical levels, operational ranges, product forms, and business models. They cannot be simply categorized as the same type of commercial vehicle ADAS product.
What they truly share is that companies are trying to find customers and business scenarios outside of passenger vehicles.
04. Technology Can Be Transferred, But Commercial Vehicles Are Not Simply Enlarged Passenger Vehicles
Cross-domain expansion by suppliers seems to have a natural logical reason.
Passenger vehicle ADAS has already accumulated massive amounts of data, models, and engineering experience. Wouldn't transferring these capabilities to commercial vehicles reduce R&D costs and shorten development cycles?
This direction is reasonable, but the practical difficulty is far greater than imagined.
There are obvious differences between passenger and commercial vehicles in terms of structure, mass, wheelbase, steering characteristics, braking systems, and operating conditions.
Taking heavy-duty trucks as an example, the mass difference between fully loaded and empty states is significant, and braking distance and vehicle response change with load.
When turning, heavy-duty trucks require consideration of larger inner wheel offset and external swing ranges. When encountering crosswinds, long downhill slopes, or continuous curves on highways, the control system faces dynamic constraints different from those of passenger vehicles.
These differences directly affect the decision-making and control of ADAS systems.
After identifying leading vehicles, lane lines, and road boundaries, passenger vehicle systems generate control commands based on the vehicle's own steering and braking capabilities. For heavy-duty trucks, the same trajectory and control commands may not yield the same vehicle response.
Therefore, while some perception, prediction, and planning algorithms have a basis for reuse, vehicle dynamic models, trajectory tracking control, braking response calibration, safety strategies, and whole-vehicle validation need to be re-adapted for specific models.
Products involving L4 autonomous driving also need to further consider steer-by-wire chassis, redundancy in key execution systems, and fault degradation capabilities.
In addition, data cannot be simply mixed.
Large amounts of driving data accumulated by passenger vehicles can help models understand road structures, traffic participants, and common driving behaviors, but heavy-duty trucks face unique challenges regarding vehicle dimensions, blind spots, loads, road access restrictions, and transport scenarios.
Sensor mounting positions, angles, perception target scales, and vehicle motion states also differ. For instance, for the same leading vehicle, the occlusion relationships, available following distance, and safe gaps for lane changing are different for passenger vehicles versus heavy-duty trucks.
Especially in high-speed trunk transportation, the system not only needs to identify leading vehicles but also judge safe distances from large vehicles, control speed on long downhill slopes, manage vehicle interactions during ramp merging, and assess braking capabilities under different load states.
Low-speed unmanned delivery vehicles present yet another set of problems.
Although they operate at lower speeds, they must deal with complex situations involving non-motorized vehicles, pedestrians, roadside parking, temporary obstacles, and narrow roads. They also need to consider loading/unloading, parking spot selection, vehicle scheduling, and last-mile delivery.
Thus, the technical accumulation in passenger vehicle ADAS can serve as a foundation for entering new markets, but it cannot be directly equated with product competitiveness in new scenarios.
What truly determines whether suppliers can stand firm is their ability to combine existing algorithmic capabilities with new vehicle models, new scenarios, and new operational systems.
This also explains why some companies choose to cooperate with commercial vehicle OEMs and logistics platforms, rather than entering the market alone with a standalone technical system.
The implementation of commercial vehicle ADAS requires OEMs to provide vehicle platforms and engineering capabilities, logistics companies to provide real operational scenarios, and operators to handle vehicle management, maintenance, and scheduling.
It is difficult for a single algorithm company to complete the entire business loop on its own.
05. Passenger Vehicles Are Not the End Point, and Commercial Vehicles Are Not a New Safe Haven
From Zhuoyu launching its commercial vehicle brand, to Pony.ai advancing the mass production of autonomous heavy-duty trucks, to QCraft, Minieye, and Desay SV Automotive entering the unmanned logistics field, the cross-scenario expansion by ADAS companies has formed a noticeable trend.
However, this does not mean that the commercial vehicle market has become a blue ocean.
First, commercial vehicle customers' procurement logic places greater emphasis on actual returns.
Whether a technical solution is advanced does not necessarily translate directly into orders. Whether the system can run stably long-term, reduce operating costs, and how faults are handled will all affect customers' willingness to make sustained purchases.
For logistics companies, while saving fuel costs and reducing labor expenses are important, factors such as whether vehicles require additional modifications, whether attendance rates are affected, and whether downtime occurs after failures also influence final procurement decisions.
Commercial vehicle customers not only focus on how much benefit can be generated but also calculate input costs and payback periods.
Second, the mass-production cycles and verification requirements for commercial vehicle projects are not low.
Vehicles operate for long hours with high intensity, requiring systems to withstand long-term road running and various working conditions. Even if a design-in is secured or the first batch is delivered, it does not mean scaled commercial revenue has been formed.
Third, business models cannot be fully shared across different commercial scenarios.
Heavy-duty truck assistance driving demonstrates value through reduced fuel consumption and driving burden; L4 heavy-duty trucks need further verification of unmanned transport operational efficiency and safety management capabilities; last-mile unmanned delivery must address issues such as vehicle costs, delivery density, scheduling efficiency, and road management.
All these markets offer opportunities, but each requires independent verification.
Therefore, cross-domain expansion by suppliers is not simply moving from a fiercely competitive market to an easier-to-profit-from one, but rather seeking new scenarios where technical capabilities can generate actual commercial value.
For passenger vehicle ADAS suppliers, the past competition was about who could be the first to implement a certain function or secure more model design-ins.
In the future, competition may further differentiate.
Some companies will continue to deepen their roots in the passenger vehicle market, building advantages through scaled mass production and cost control; others will extend their technology platforms to heavy-duty trucks and logistics; still others will continue to expand into L4 autonomous driving and unmanned operations.
Passenger vehicle ADAS certainly has not lost its appeal, but it is no longer the only direction all ADAS companies need to bet on.
As more and more suppliers begin to seek markets outside of passenger vehicles, what is truly changing in the industry may not be that ADAS technology is moving from passenger vehicles to commercial vehicles, but that companies are beginning to rethink who ADAS technology should ultimately be sold to, and through what methods it should create sustainable, redeemable commercial value.