Accelerated Deployment of Intelligent Driving: L3 Regulations Clear Hurdles, L4 Leads in Logistics and Mobility

As regulations mature and business models gain validation, 2026 is shaping up to be a watershed year for advanced autonomous driving, marking the shift from technological breakthroughs to mass production.

On March 2, industry observer Gasgoo noted that Joyson Electronics announced a significant breakthrough in advanced autonomous driving. The company plans to begin mass production of its first L3-level intelligent driving domain controller by mid-2027, which will be featured in models from a leading automaker. Simultaneously, its L4 domain controller is set to be deployed in low-speed unmanned logistics vehicles, potentially becoming the first mass-produced L4 controller based on a domestic chip platform.

This progress represents more than just a technological win for a single company; it is a microcosm of the broader industrial ecosystem reaching maturity.

Driven by the dual forces of falling regulatory standards and a closing commercial loop, China’s autonomous driving sector is shifting into high gear.

The Institutional Foundation for L3 Autonomous Driving Is Solidified

The reason 2026 is viewed as the inaugural year for L3 mass production lies fundamentally in the critical policy groundwork laid at the government level.

Effective January 1, several mandatory national standards—including the "Technical Requirements for Automobile Information Security" and "General Technical Requirements for Automotive Software Upgrades"—have taken effect, establishing a clear compliance framework for the legal road use of advanced autonomous driving systems.

Specifically for L3 autonomous driving, the new regulations clarify the mechanism for liability transfer at the legal level for the first time: if the system is active, no takeover request has been issued, and an accident occurs, the vehicle manufacturer bears the responsibility according to the law.

The implementation of this clause resolves the long-standing challenge of defining rights and responsibilities that has constrained L3 development.

At the same time, mandatory standards impose stringent requirements on technical details.

All L3 vehicles must be equipped with "black box" data recording systems to log system activation status, takeover request times, and driver response behavior in real time. Automakers are required to purchase specialized autonomous driving liability insurance of no less than 5 million yuan. Furthermore, systems must continuously monitor whether the driver is capable of taking over using at least two valid indicators.

These regulations steer the industry away from its previous "function-over-safety" bias and back onto a safety-centric track. It also implies that only companies possessing full-stack R&D capabilities and robust compliance systems will be able to gain a foothold in the market.

Joyson Electronics’ disclosed timeline for mass-producing its L3 domain controller lands squarely within this institutional window.

Its ability to achieve mass production by mid-2027 is underpinned by early adaptation to these new compliance requirements. Industry trends indicate that the penetration rate of mapless autonomous driving is expected to reach 8% in 2026, while the sales share of L3 models could hit 15% by 2028.

As the cost proportion of LiDAR and advanced autonomous driving systems gradually declines, L3 autonomous driving is moving from a feature in a few high-end models to a much broader market space.

L4 Autonomous Driving Achieves Commercial Viability in Niche Sectors First

Unlike the progressive development path of L3, L4 autonomous driving follows a pragmatic route defined by "segmented scenarios and a focus on operations."

In 2026, scenario-based L4 autonomous driving is reaching a breakthrough inflection point from zero to one. Large-scale applications have already launched, particularly in commercial scenarios such as freight, operations, and closed-campus roads.

The L4 domain controller recently launched by Joyson Electronics is set to be deployed in low-speed unmanned logistics vehicles and is poised to become the first mass-produced L4 product based on a domestic chip platform.

The industry context for this move is that unmanned logistics vehicles are the first to successfully validate business models, thanks to their unique technical economics.

As of June 2025, 103 cities in China had launched pilot programs granting road rights to unmanned logistics vehicles, covering over 80% of major logistics hub cities. The characteristics of low-speed operation, fixed routes, and controllable scenarios mean these vehicles have relatively moderate demands for hardware computing power. This allows for the use of more cost-effective sensor combinations, thereby achieving a superior economic model.

In terms of passenger scenarios, the commercial validation of Robotaxis has also achieved key breakthroughs.

Pony.ai's seventh-generation Robotaxi achieved positive monthly operating profit per vehicle in Shenzhen this February. Daily orders per vehicle reached 23, with average daily net income hitting 338 yuan.

Following its achievement of this target in Guangzhou last November, Pony.ai has now proven in two first-tier cities that autonomous mobility is not only technically feasible but also economically sustainable at scale. By mid-February 2026, its paid orders in Shenzhen for the year had already surpassed the total for all of 2025.

Furthermore, in higher-level L4 scenarios, Joyson Electronics and SiNian are jointly expanding L4 scale applications. A smart port digital management platform they co-developed—based on "V2X + L4 autonomous driving + intelligent cloud dispatch"—has been deployed and is operating stably at ports including Ningbo and Xiamen.

Image Source: Joyson Electronics

This multi-scenario layout—spanning port logistics, last-mile delivery, and urban Robotaxis—reflects a shift in positioning for Tier 1 suppliers in the era of advanced autonomous driving. They are evolving from mere hardware providers into system solution integrators, covering domain controllers and cross-domain converged central computing units.

Judging by Joyson Electronics' performance in the third quarter of 2025 alone—securing over 20 billion yuan in intelligent orders—market expectations for advanced autonomous driving are rapidly being realized.

With regulatory hurdles established and business models validated, 2026 is undeniably a historic milestone for the industrialization of high-level autonomous driving. Companies that can implement technology within compliance frameworks and close commercial loops in niche sectors will seize the first-mover advantage in the impending boom.