Zhuoyu Takes the Lead as the Window for Integrated Cockpit-Driving Solutions Narrows

The concept of integrated cockpit-driving systems has been touted for years, yet truly mass-produced solutions remain few and far between.

On May 27, Zhuoyu Technology delivered its third solution with the Arcfox Wendao V9, a full-size MPV starting at 194,800 yuan. Seven months prior, the Alpha T5 and Alpha S5 had already been launched. With SUVs, sedans, and MPVs now all covered, Zhuoyu has successfully implemented its solution across three body types. During the same period, most other integrated cockpit-driving solutions were still stuck at the nomination or engineering validation stages.

Zhuoyu employs the Qualcomm SA8775P chip. In an era where intelligent driving chips often boast computing powers exceeding 1,000 TOPS, its performance is not particularly high. Its strategy is not about stacking parameters, but rather repeatedly utilizing a single architecture. By platformizing the underlying software stack, it achieves rapid adaptation for different wheelbases and powertrain types, avoiding the need to start from scratch when moving from the T5 to the S5 and then to the V9. Each additional mass-produced vehicle adds another card to the table for future nomination negotiations.

In Zhuoyu's solution, the cockpit and intelligent driving systems run on two isolated operating systems. The intelligent driving system operates in an independent secure environment, unaffected by UI rendering interference. Computing power is dynamically allocated via software: it is lent to the cockpit when parked and prioritized for intelligent driving while the vehicle is in motion. Liquid cooling design keeps temperatures in check under full load, with over 30% of computing power reserved for subsequent OTA updates. For users, the most intuitive perception is response speed. Saying "start parking" instantly brings up the interface, and switching songs or adjusting the air conditioning during intelligent driving does not cause lag. This is because direct data connections between the cockpit and driving systems have reduced communication latency from milliseconds to microseconds.

Architecture is merely the foundation; the perception and decision-making layers running on top determine the user experience. Zhuoyu's inertial navigation binocular system obtains depth information through stereoscopic vision parallax, generating dense point clouds. It can identify and avoid irregular obstacles such as traffic cones, water-filled barriers, and rocks, maintaining stable performance even in rain, fog, or at night. The decision-making layer utilizes a highly perceptive end-to-end model trained with reinforcement learning.

Platform-based reuse constitutes another layer of competitiveness. The underlying architecture and software stack do not require rebuilding for different vehicle models, allowing for rapid adaptation across varying wheelbases and powertrain types.

Suppliers vying for this track are looking at a very specific financial calculation.

Yu Kai, founder of Horizon Robotics, previously stated to Wallstreetcn that single-chip solutions reduce hardware costs by 20% to 30%, saving 1,500 to 4,000 yuan per vehicle, and compressing the R&D and delivery cycle from 18 months to 8 months.

According to statistics from Zozy Auto Research, sales of mass-produced vehicles with integrated cockpit-driving systems will reach 1.67 million units in 2025, a year-on-year increase of 43%. However, only a minority have successfully implemented "integrated cockpit-driving" single-chip solutions.

The true significance of this calculation lies in the mid-to-low-end market. The 100,000 to 200,000 yuan price range is the mainstay of the Chinese auto market. Urban NOA (Navigate on Autopilot) remains a scarce configuration in this segment, not due algorithmic limitations, but because the cost of two separate domain controllers is unsustainable. The BOM savings from integrated cockpit-driving systems are sufficient for automakers to make advanced intelligent driving standard equipment rather than an optional extra. The Wendao V9 is a prime example: an MPV starting at 194,800 yuan comes with urban NOA and full-scenario automatic parking as standard features across all trims. Two years ago, such configurations were only available on vehicles priced around 300,000 yuan. Whoever helps automakers achieve this first wins the nominations.

This track is becoming increasingly crowded. Within the Qualcomm 8775 camp, Desay SV's integrated cockpit-driving domain controller has reduced the combined cost of smart cockpit and smart driving to below 6,000 yuan, securing nominations from Chery and Tata. Bosch is simultaneously adapting to both Qualcomm and MediaTek lines. At the Beijing Auto Show, CarConnect, Huayang, and ThunderSoft collectively showcased their 8775 solutions. Outside this camp, Horizon Robotics' Starry chip, built on a 5nm process with 650 TOPS, made its debut with iCAR. Black Sesame's Wudang C1200 is tied to Dongfeng. Qualcomm's own SA8797 boosts computing power to 320 TOPS, with Li Auto and Leapmotor already signed up.

Before the 8775 solution has firmly established itself, the pressure to iterate to the next generation of chips has already arrived.

However, computing power is not the only barrier. At the Beijing Auto Show, some launched vehicles with integrated cockpit-driving systems faced issues with inflated computing power claims and experiential lag. Engineering pitfalls such as virtualization isolation, computing power scheduling, and thermal management can only be resolved through repeated mass production. Zhuoyu's experience gained across three vehicle types is indeed scarce in the short term.

Data from Zhuoyu shows that as of March 2026, it has 20 cooperative clients covering 32 brands, with cumulative mass-produced models exceeding 50 and nominations surpassing 100. Its client list covers half of the Chinese market. Integrated cockpit-driving is just one line in Zhuoyu's current expansion path, but it may be the most critical.

The shelf life of experience is also limited. As more solution providers complete their initial mass production and higher-computing-power chips level the engineering playing field, the moat for early movers will shallow. Zozy predicts a compound annual growth rate of 36% for this market from 2026 to 2030, indicating significant incremental space, but seats on the track are limited. The pace of mass production in the next year or two will likely determine how long each supplier can remain at the table.

Risk Warning and Disclaimer

The market carries risks, and investment requires caution. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial status, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investment based on this content is at the user's own risk.