Under the new OTA regulations, how can car manufacturers adapt to the new iteration model?

When automobile enterprises no longer solely pursue "iteration speed" but start to build "safety depth", the competition shifts from flaunting functions to a deeper confrontation in system resilience, user trust, and ethical responsibility.

"The intelligent driving industry for automobiles is highly competitive. We provide technologies to help everyone compete in this field." When Yu Chengdong said this, it was an aggressive era when Chinese automobile enterprises were continuously developing "cool functions" and focusing on "monthly OTA (Over-the-Air remote vehicle upgrade)".

However, on the other side of the upgrade competition, data from the State Administration for Market Regulation shows that in 2024, the number of vehicle recalls in China increased by 67% year-on-year, and the recall situation of new energy vehicles was particularly prominent. This has also forced Chinese automobile enterprises to calm down and shift their R & D focus more towards "safety redundancy".

Regulatory authorities have clearly noticed this industry situation. Not long ago, the State Administration for Market Regulation (referred to as "SAMR") and the Ministry of Industry and Information Technology jointly issued the "Notice (Draft for Comments) on Strengthening the Supervision and Management of Product Recalls and Production Consistency and Standardizing the Promotion of Intelligent Connected New Energy Vehicles", soliciting public opinions.

From the current information in the draft for comments, it covers four major aspects, including vehicle recalls, production consistency, commercial promotion, and incident and accident reporting, all of which are the most concerned issues for consumers of intelligent connected new energy vehicles at present. In particular, it restricts the push of software versions that have not been fully tested and verified to users and the concealment of defects through OTA.

Yang Dan, a practitioner in the new energy field, frankly told Lu Jiu Business Review that this is actually a positive thing and an opportunity to rebuild trust. After all, experts from relevant ministries and commissions have been emphasizing that "when a car becomes a smart terminal on four wheels, its safety standards should be comparable to those of the civil aviation industry."

01 The Reconstruction of Compliance May Bring a Painful Period

The introduction of the new OTA regulations is equivalent to forcibly attaching a "safety certificate" to all the "cool technologies" packaging. In the past two years, the high - frequency OTA upgrades of many automobile brands have made it difficult for users to make judgments and choices. Users also have questions: Do I really need all these?

In response, Zeng Chun, the editor - in - chief of "China Industry and Information Technology", told Lu Jiu Business Review: "As we have observed, OTA is closely related to vehicle intelligence, and its core value lies in achieving continuous evolution of functions."

In his view, since the R & D cycle of a whole vehicle is long, usually taking several years, while the iteration of electronic hardware technology is extremely fast, the industry generally adopts the "hardware first" strategy - that is, deploying high - performance hardware in advance during the vehicle design stage. This can ensure that the vehicle's hardware foundation has long - term competitiveness. More importantly, with the help of OTA technology, automakers can gradually release new functions based on technological progress and user feedback, extending the vehicle's life cycle.

"This combination of the 'hardware first' approach and OTA also reserves technical space for the future upgrade of the Advanced Driver - Assistance System (ADAS). The newly introduced OTA regulations are precisely to standardize and guide this iterative process to make it develop more safely and orderly," Zeng Chun added.

Here, we have to mention XPeng, one of the representatives of new - force automakers. Before Huawei entered the intelligent driving industry for automobiles, Yu Chengdong also publicly stated that "XPeng is the best in intelligent driving in China." In 2022, XPeng Motors established a special OTA committee during a major organizational structure adjustment.

Not long ago, the XPeng MONA M03 underwent its ninth large - scale OTA upgrade. This upgrade included 30 new features and optimizations, covering multiple aspects such as the AI intelligent cockpit, navigation system, intelligent voice assistant Xiao P, charging report, and driving assistance system (such as the improvement of AEB performance). This update frequency and scope reflect to some extent its previous "agile development" model.

With the strengthening of OTA upgrade supervision by relevant departments (such as banning "mid - night pushes" and requiring pre - upgrade filings), for automakers like XPeng, the entire OTA R & D, testing, and release process may be reconstructed. They must complete more rigorous internal testing, safety assessments, and filings with the competent authorities before pushing, and the cycle from "idea to launch" will inevitably be lengthened.

Enterprises like Huawei Hongmeng Zhixing will also face similar challenges as XPeng. Moreover, Huawei also has to deal with more complex "cross - terminal" issues. After all, terminal collaboration, which seamlessly connects the in - vehicle system with smart devices such as mobile phones, tablets, and watches through the Hongmeng system to achieve seamless experience transfer, is one of its core advantages.

After the introduction of the new regulations, upgrades involving changes in technical parameters need to be reported to the Ministry of Industry and Information Technology. In the future, any function update involving the vehicle's underlying technology must be coordinated with supply - chain partners and filed. The "agility" and cross - terminal experience consistency that the Hongmeng cockpit is proud of may face more complex process tests.

In terms of safety redundancy, before the introduction of the new regulations, the warning strategy of the in - vehicle system for driver monitoring (DMS) was relatively simple, perhaps only limited to steering wheel grip detection and simple warning sounds. Once the new regulations are implemented, "three - level warning until forced parking" is a clear and progressive mandatory safety redundancy.

The new regulations mandate the introduction of a "three - level alarm" mechanism, which requires the development of a complete redundancy logic from sensory warnings to forced parking and multiple verifications at the algorithm level of perception and decision - making, significantly increasing the computing power load. According to industry estimates, the relevant hardware and R & D costs will increase by about 20%. The resulting cost pressure will also increase accordingly.

The regulatory new regulations require the establishment of an "intelligent vehicle black - box" system, which can report intelligent driving accidents in real - time and quickly trace the source. This is more beneficial for enterprises like Huawei, which has decades of technological accumulation in the ICT field. For example, in the Wenjie model equipped with Huawei ADS 2.0, the system can automatically encrypt and upload desensitized data, and the cloud platform can quickly simulate and reproduce potential dangerous scenarios to accurately locate the problem version and its scope of influence.

In contrast, many new - force automakers still manage data in traditional ways, and it is difficult to integrate data from different departments (such as R & D, after - sales, and quality). Facing sudden "spot checks", they may need to coordinate data from multiple parties temporarily, which is inefficient and error - prone. Therefore, the data closed - loop ability required by the new regulations is the strength of technology companies like Huawei, and it is also a basic course that many new - force automakers must make up for.

02 "Vehicle Iteration" Will Enter the Standardization Era

People in the industry generally believe that, in addition to providing clearer standards at the regulatory level, the biggest change brought by the new OTA regulations to the intelligent driving industry may be the shaking of the "hardware pre - installation + software unlocking" model that has been maintained for many years.

Zhao Xinlei, a senior technical expert from a new - force automobile brand, shared his views with Lu Jiu Business Review: "Upgrading vehicle hardware through remote software updates involves two key links: one is the pre - installation of hardware during the vehicle manufacturing process, and the other is the software upgrade during the vehicle service process. Tesla has been able to achieve multiple OTA upgrades, mostly relying on the original hardware pre - installation. The so - called 'upgrade' is actually the 'unlocking' of some advanced functions."

A typical example is Tesla. From HW 1.0 to the current HW 4.0, each generation of Tesla's autonomous driving hardware (HW) has pre - installed powerful computing power and sensor interfaces. Tesla then uses numerous OTAs over several years to gradually unlock and optimize its intelligent driving capabilities.

Zhao Xinlei mentioned that "this is the common practice in the industry. Of course, this directly affects the development logic of automobile products. For example, to upgrade intelligent driving functions, it is necessary to pre - install enough cameras and lidars in advance. When you buy a car, you have already paid for these pre - installed hardware. Accepting subsequent OTA upgrades and paying a certain fee also helps the manufacturer share the R & D costs."

To date, it is very clear that the regulatory new regulations are shaking the traditional business model of intelligent driving for automobiles. Leading enterprises are trying to adapt to the new iteration model based on their own advantages.

For example, XPeng, which has a first - mover advantage in the OTA field, its increasingly mature "automatic roll - back mechanism" is no longer just a simple version restoration. XPeng divides the in - vehicle system into multiple independent functional domains (such as intelligent driving, cockpit, and power). If any module fails to upgrade, it can be independently rolled back without affecting the operation of the whole vehicle.

This mechanism was verified in the latest OTA of the MONA M03. When some users' upgrades were interrupted due to network fluctuations, the system automatically recognized and restored to a stable version within 90 seconds without causing any vehicle - using failures. Against the background of the new regulations emphasizing "zero - risk upgrades", XPeng Motors has begun to clearly mark the roll - back trigger conditions and processes in the user manual.

Ideal's safety redundancy was once regarded as "over - design", but now it exactly matches the requirements of "three - level warning - forced parking" in the new regulations. For example, when the system detects that the driver is continuously distracted and the DMS alarm reaches the second level, the redundant controller can immediately take over the vehicle when the main system fails, gradually reduce the speed, turn on the hazard lights, and finally achieve the third - level forced parking.

Taking the latest L8 as an example, in its marketing language, Ideal also discloses the "safety redundancy rate" as a core parameter (such as the braking redundancy response < 100ms) and compares it with the aviation - level backup standard. This has also become a key support for its brand premium in the market above 200,000 yuan.

As mentioned above about the synergy effect, Huawei Hongmeng Zhixing may continue to focus on the "holistic thinking". The core safety logic of the vehicle, such as braking and warning, is very stable. In addition to integrating multiple heterogeneous sensors such as cameras, millimeter - wave radars, ultrasonic radars, and lidars to achieve higher - level environmental perception, it also has an anti - misoperation mechanism. It will not hand over control until it eliminates the driver's nervous misjudgment, and this will not be changed usually.

So each upgrade is to optimize the experience and expand functions on the basis of safety. It does not upgrade the car or mobile phone separately but "packages" devices such as mobile phones, watches, tablets, and in - vehicle systems as a whole for upgrade and filing. When the mobile phone and in - vehicle system are jointly upgraded to HarmonyOS 5.0, the "cross - device dangerous scenario warning" function can be enabled.

When automobile enterprises no longer solely pursue "iteration speed" but start to build "safety depth", the competition shifts from flaunting functions to a deeper confrontation in system resilience, user trust, and ethical responsibility. Reshaping the evolution logic under constraints is actually a response to the regulatory trend of "upgrading responsibility".

This article is from the WeChat public account "Lu Jiu Finance" (ID: liujiucaijing69), author: Zhou Qian, published by 36Kr with authorization.