Huawei on Release of Mandatory National Standard for Autonomous Driving Systems; Yinwang among First to Complete L3 Vehicle Access Pilot Verification
Core Highlights
On July 30, 2026, the Ministry of Industry and Information Technology (MIIT) officially released GB 44721-2026, "Intelligent and Connected Vehicles — Safety Requirements for Automated Driving Systems." This is China's first mandatory national standard specifically targeting L3 and L4 automated driving systems, and its publication is a landmark moment for the domestic intelligent-driving industry. It marks the point at which autonomous driving moves out of the realm of recommended and guidance-level specifications and into an era of hard constraints: if an automaker wishes to mass-produce high-level driver-assistance, it must first clear this compliance threshold. In statements released by Huawei, its subsidiary Yinwang (Huawei Intelligent Automotive) was named among the first batch of enterprises in China to complete L3 vehicle access pilot verification. This indirectly confirms Yinwang's first-mover advantage in the compliance of intelligent driving. For the wider industry, the weight of this standard is comparable to a "birth certificate" that grants access to the nationwide rollout of high-level intelligent driving, and it signals that the regulatory environment around self-driving in China is finally maturing into something concrete and enforceable. The timing is especially significant because it lands as Chinese automakers accelerate commercialization of advanced assisted driving to stay ahead of foreign competitors.
What Happened
The standard lays out safety requirements across the entire chain of perception, decision-making, and execution for automated driving systems, covering core areas such as dynamic driving task take-over, failure-response for human-machine interaction breakdowns, data storage, and event traceability. Prior to this, China's L3 pilots had long existed in a state of "regionally permitted but without a unified national yardstick," in which the mutual recognition of tests across different cities carried high costs and forced automakers into repeated investment. After the release of the mandatory national standard, the implementation date of July 1, 2027 becomes a clear watershed: products that fail to meet the requirements cannot enter the announcement catalog, and therefore cannot be legally sold or driven on public roads. Huawei Yinwang's early verification is equivalent to seizing a compliant position ahead of the policy's enforcement, clearing the institutional obstacles for the subsequent mass production of its vehicle models and giving it a measurable head start over rivals who have not yet completed the process.
Technical Details
The central difficulty of the standard lies in defining the "boundary of responsibility." L3 is conditional automated driving: when the system is active, the responsible party shifts from the driver to the automaker, which is why the standard imposes strict metrics on the system's fail-operational capability, redundant braking and steering, and driver state monitoring. As the core intelligent-vehicle entity spun out from Huawei's car BU, Yinwang's solution naturally inherits Huawei's accumulated strengths in redundant architecture and vehicle-cloud coordination, two areas where Huawei has invested heavily for years. This is precisely the technical confidence that allowed it to pass pilot verification so quickly. It is also worth noting that the standard strengthens requirements for data security and event recording, so that every intelligent-driving take-over can be retrospectively traced, an element that is essential for liability determination after incidents and for continuous safety improvement. Such requirements also push suppliers to standardize their interfaces, which over time lowers integration cost across the whole vehicle supply chain.
Comparison with Competitors
Looking horizontally, the EU's UN-R157 also legislates for L3 automated lane-keeping systems, but China is the first country to extend coverage to L4 in the form of a mandatory national standard. Compared with industry-group standards drafted by automakers banding together, the national mandatory standard unifies the testing caliber and lowers the compliance trial-and-error cost for small and mid-sized automakers. At the same time, it raises the overall industry threshold, which may accelerate the elimination of trailing players who lack the engineering resources to meet the bar. For leading manufacturers that already possess full-stack in-house R&D capability, this tightening instead becomes a moat that protects market share and raises the cost of entry for latecomers who might otherwise undercut on price.
Industry Impact and Use Cases
For consumers, the mandatory national standard delivers a trustworthy safety floor rather than marketing rhetoric, and it gives buyers a concrete basis for comparing the maturity of different brands' assisted-driving offerings. For the industry, the policy window is now explicit: the period from 2026 to 2027 is the critical year in which automakers must sprint toward access approval. By seizing the compliance first-mover advantage, Huawei Yinwang is well positioned to expand its voice in intelligent-driving licensing and component supply, reinforcing its underlying strategic logic of "we do not build cars, we help automakers build good cars." From a broader perspective, a unified standard also helps Chinese intelligent-vehicle solutions gain a foundation for international mutual recognition when they expand overseas, reducing the friction of entering foreign markets and strengthening the global competitiveness of the domestic supply chain. It equally gives Chinese regulators a proven template that may shape how other emerging technology domains are standardized in the coming years.