Gasgoo Munich-As automakers race to accelerate R&D, regulators are drawing a clearer safety line for new model launches.
On October 10, the Ministry of Industry and Information Technology (MIIT) and three other departments released a draft notice seeking public comment on strengthening the management of automotive product innovation, design, and R&D testing verification. The proposal seeks to ban fully hidden door handles, mandate physical controls for critical functions like lights and wipers, and prohibit the use of "zero-gravity" seats while driving. More notably, new vehicle reliability verification tests must cover at least 30,000 kilometers, while environmental adaptation verification must span at least one year, covering "two winters and one summer."
AI-generated image
Just a month ago at the 4th AI-Defined Vehicle Forum, Wang Xianbin, vice president of Gasgoo Automotive Research Institute, highlighted the pressure a 12-month development cycle places on the supply chain: As vehicle development cycles shrink, some suppliers are left with just three to four months to complete R&D and verification.
Vehicle development may be accelerating, but how much room is left to compress engineering verification when safety is on the line?
Hyped Innovations Are Facing a Reckoning
Over the past few years, a wave of new designs has flooded the new-energy vehicle (NEV) market.
Hidden door handles, yoke steering wheels, the removal of physical buttons, zero-gravity seats, and external light signals—these configurations have become crucial tools for automakers to showcase technology and differentiation. Yet as their application scales up, safety concerns lurking behind some of these designs are beginning to surface.
Hidden door handles are a prime example.
Originally serving aerodynamics and styling needs, the feature gradually became a signature of NEVs. However, when faced with low-temperature icing, electronic mechanism failure, or power loss after a collision, whether a door can be opened promptly is no longer just a matter of convenience.
On January 28, 2026, the State Administration for Market Regulation and the Standardization Administration of China approved and released the "Safety Technical Requirements for Automobile Door Handles" (GB 48001—2026). Set to take effect on January 1, 2027, it establishes clear requirements for the mechanical opening capability of door handles.
The latest draft goes a step further, proposing a ban on fully hidden door handles and requiring that innovative designs include mechanical redundancy and adequate hand-operating space.
Similar shifts are occurring inside the cabin.
As more automakers integrate functions like lights and wipers into central control screens—making minimalist interiors a trend—whether drivers can operate them quickly and accurately in emergencies is easily overlooked.
Consequently, the draft mandates that even if touch controls are used for these driving-safety functions, physical controls must be retained. For features like zero-gravity seats, rotating seats, and reclining seats that convert into beds, use is prohibited while the vehicle is in motion.
Regulators are not rejecting automotive innovation outright; rather, they are requiring automakers to fully consider safety risks under abnormal operating conditions while pursuing functional and experiential differentiation.
Just because a design can be realized doesn't mean it's ready for mass production. And just because consumers are willing to pay for new features doesn't mean those features have been fully verified.
Building a Car in 12 Months: Where Does the Verification Time Come From?
Compared to the tweaks on popular features, the draft's requirements for R&D testing could have far more profound implications.
Shortening development cycles has become a key competitive weapon for Chinese automakers in recent years. Traditional development often took several years, but some companies have now compressed that to around 18 months, or even attempted to slash it further to 12.
This shift is underpinned by technological progress.
At the same forum, Meng Chao, senior director of SAIC Motor's Intelligent Software Center and CTO of Z-one Tech, noted that through software architecture optimization, hardware-software decoupling, and OTA technology, vehicle development cycles can be shortened to 18 to 20 months, while software development cycles can drop to just 6 to 8 months.
Platform-based development, the reuse of mature components, digital simulation, and parallel development can indeed reduce repetitive work and boost efficiency.
But certain stages cannot be compressed at the same rate.
Chassis systems must endure repeated impacts across different road conditions, power batteries need to withstand extreme heat and cold, and electronic-electrical systems must verify long-term stability in complex environments. Laboratories can simulate some conditions, but they can hardly fully replace the complex environment of real-world roads.
For suppliers, the pressure is even more immediate.
A new vehicle involves multiple systems. From plan finalization and sample development to vehicle integration, every step has its own timeline. If vehicle project milestones are constantly moved up, suppliers are forced to complete development, testing, and rectification in less time.
Wang Xianbin put it bluntly: "If you still miss the deadline after cutting everything else, the only knife left to use is on verification."
This draft notice aims precisely to hold that line.
According to the proposed requirements, new vehicle reliability verification tests must cover no less than 30,000 kilometers. Environmental adaptation verification must span at least one year, covering "two winters and one summer"—a cycle of extreme cold, extreme heat, and extreme cold again.
The document permits off-season testing and environmental chamber simulation, provided the verification results are no less effective than actual road tests. For complex environments that are difficult to replicate, real-vehicle testing remains mandatory.
Crucially, if relevant R&D verification is incomplete or safety hazards found during testing remain unresolved, the model cannot be submitted for product approval.
These requirements are not limited to NEVs. Internal combustion engine vehicles fall under the same regulatory scope and must also complete applicable reliability, durability, and environmental adaptation verification.
Of course, a one-year environmental verification cycle does not mean the vehicle development cycle must exceed one year. By reasonably arranging parallel testing and increasing platform and component commonality, automakers still have room to improve efficiency.
The difference is that going forward, meeting launch deadlines can no longer be used as an excuse to skip verification work that should be done.
AI in Cars Makes Safety Verification Harder Than Ever
Compared to traditional mechanical structures, the challenge for intelligent vehicles lies not only in the proliferation of functions but also in the changing way some functions operate.
In the past, automotive software mostly executed tasks based on rules preset by engineers, allowing companies to test around input conditions and expected results. But with the introduction of end-to-end large models and AI agents into vehicles, the situation has become complicated.
Take assisted driving as an example. The same road scenario might lead the model to make different judgments due to subtle changes in weather, lighting, the behavior of other vehicles, or sensor input. Even if a system performs well in extensive routine testing, it doesn't guarantee it can correctly handle rare, extreme cases.
This presents a new conundrum: when a system no longer makes decisions entirely based on preset rules, how can automakers prove it is sufficiently safe?
Engineering practices at companies like FAW-Volkswagen show that intelligent driving cannot do without a massive amount of real-vehicle testing. Yet as AI model capabilities expand, simply increasing test mileage may not necessarily cover all risks.
Some truly dangerous situations may be precisely those special scenarios that rarely appear in daily driving. Therefore, automakers must not only accumulate test mileage but also actively hunt for potential system failures through simulation, scenario replay, and real-vehicle verification.
The AI infrastructure solution previously shared by Meng Chao incorporates real-world scenario data, model training, evaluation verification, and safety release into a continuous iteration loop.
This verification system is particularly critical in the OTA era. Because a single software upgrade can alter a vehicle's original control logic or system behavior, even if a previous version passed testing, the new version must be evaluated for whether it introduces other risks.
The draft explicitly requires that combined driving assistance, automated driving, and intelligent cockpit systems undergo real-vehicle testing, covering both routine usage scenarios and extreme scenarios where safety hazards may exist. For OTA upgrades, full testing and verification are equally required, along with the retention of internal corporate approval records.
Furthermore, automakers must accurately describe functional boundaries, refrain from exaggerated marketing, and certainly not induce drivers to neglect their safe driving responsibilities.
For intelligent vehicles, the launch of a feature does not mark the end of R&D verification. As software continues to update, safety assessment must span the entire lifecycle of the vehicle.
OEMs Can No Longer Simply Accept Supplier Verification Reports
If stricter testing requirements are set to impact the rhythm of automotive R&D, the further clarification of responsibility attribution could change the way OEMs and suppliers cooperate.
The draft proposes that for key hardware and software—such as driving assistance systems, intelligent cockpits, battery packs, electronic controls, and motors—OEMs must conduct incoming inspections and cannot simply rely on verification reports provided by suppliers.
At the same time, automakers must preserve original records, reports, and relevant supporting materials for R&D testing verification, with a retention period of no less than six years.
This does not imply that supplier test results lack value, but rather requires OEMs to shoulder their own quality management responsibilities.
AI-generated image
In recent years, the division of labor within China's automotive supply chain has deepened. Increasingly, Tier 1 suppliers no longer just provide individual components; they participate in vehicle function definition, software development, and even system integration.
Especially in areas like intelligent driving and brake-by-wire, the safety performance of a system often depends on the synergy between different hardware, software, and control strategies.
The fact that a single component meets performance requirements does not guarantee it will function properly under all complex conditions once installed in the vehicle.
Previous industry discussions surrounding controversies over products like braking systems have touched on this issue repeatedly: when a function malfunctions, how exactly should responsibility be distinguished between component design, software control, and vehicle integration?
For increasingly complex automotive systems, relying solely on the test results of individual components makes it difficult to fully explain a vehicle's safety performance.
Therefore, the new requirements do not ask OEMs to repeat all of a supplier's tests. Instead, they emphasize that OEMs must establish their own inspection and quality management capabilities, and cannot simply shift vehicle safety responsibility onto suppliers.
This change will also raise the engineering threshold for some automakers that rely on mature supplier solutions to rapidly integrate new vehicles.
In August of this year, the MIIT and three other departments launched a year-long special campaign on automotive product production consistency and quality improvement, focusing on issues such as reliability, durability, and the testing verification of new technologies. This latest draft further moves regulatory requirements forward to the R&D design and product approval stages.
For automakers and suppliers, stricter verification requirements will inevitably bring new costs.
Complex condition testing, the accumulation of long-term reliability data, and system-level verification all require investment in equipment, personnel, and capital. For companies with limited R&D resources that rely on rapidly launching new models to seize market share, the pressure will be even more pronounced.
On the other hand, companies that have long prioritized engineering verification and possess complete quality management systems may gain a new competitive advantage.
In the future, automotive suppliers will compete not only on performance, price, and delivery speed, but also on their track record of reliability and their ability to support OEMs in completing system-level verification.
For OEMs, R&D efficiency cannot be measured solely by development cycle time. Even if a new vehicle launches early, if it subsequently suffers frequent quality issues or requires recalls, the time saved upfront may not have created real value.
According to the draft's proposed schedule, starting January 1, 2027, newly declared models involving these requirements must submit supplementary technical parameters and verification materials. Relevant models that have already obtained announcement approval must complete supplementary filings by July 1, 2027. The document is currently in the consultation phase, and final regulations are yet to be determined.
But the regulatory direction is clear.
The speed advantage established by China's automotive industry in recent years through supply chain collaboration, technological innovation, and improved R&D efficiency does not need to be negated.
What truly warrants caution is when market competition turns speed into the most important metric, and safety verification is treated as a cost that can be endlessly compressed.
Cars may iterate as fast as consumer electronics, but they cannot be developed using the fault-tolerance logic of consumer electronics.
In the next stage, what China's automotive industry needs to prove is not just whether it can build a car in 12 months, but whether it can still complete thorough verification within an even shorter development cycle.
After all, building cars can get faster—but some time cannot be saved.
