卓越驱动抢先起步，舱驾一体化的窗口期正在收窄。

Integrated cockpit-driving systems have been talked about for years, but truly mass-produced solutions that make it into vehicles are few and far between.

On May 27th, Zhuoyu Technology launched its third model, the Arcfox Wen Dao V9, a full-sized MPV starting at 194,800 yuan. Looking back seven months, the Alpha T5 and Alpha S5 were already released. With SUVs, sedans, and MPVs, all three body types have been implemented. During this period, most integrated cockpit-driving solutions are still stuck at the fixed-point or engineering validation stage.

Zhuoyu uses Qualcomm’s SA8775P. In today’s era where smart driving chips often boast thousands of TOPS of computing power, its performance isn’t particularly high. Its strategy isn’t about stacking parameters, but reusing the same architecture repeatedly, with the underlying software stack platformized for fast adaptation to different wheelbases and power types. T5 to S5 to V9—no need to start from scratch each time. The more models are mass-produced, the more bargaining chips Zhuoyu has for future fixed-point negotiations.

In Zhuoyu’s solution, the cockpit and smart driving run on two separate isolated operating systems; smart driving operates in an independent safety environment, unaffected by UI rendering. Computing power is dynamically allocated via software, lent to the cockpit when parked, with smart driving prioritized when driving. Liquid cooling keeps temperatures under load in check, reserving over 30% computing power for future OTA upgrades. For users, the most direct experience is the response speed: when you say “start parking,” the interface instantly pops up; adjusting music or air conditioning during smart driving doesn’t lag. Behind this is the direct data connection between the cockpit and driving subsystems, dropping communication latency from milliseconds to microseconds.

The architecture is just the foundation; perception and decision-making running on it determine the real experience. Zhuoyu’s inertial navigation stereo vision uses depth information from parallax to generate dense point clouds, allowing it to recognize and avoid irregular obstacles like traffic cones, barriers, or stones, working reliably in rain, fog, or nighttime. The decision layer is a highly adaptive end-to-end model trained with reinforcement learning.

Platformized reuse is another layer of competitiveness. The underlying architecture and software stack aren’t rebuilt for each vehicle model, enabling fast adaptation to different wheelbases and power types.

Suppliers are competing for this track, driven by very specific calculations.

Yu Kai, founder of Horizon Robotics, previously told Wall Street Insights: A single-chip solution cuts hardware costs by 20%-30%, saving 1,500–4,000 yuan per vehicle, and reduces the R&D and delivery cycle from 18 months to 8 months.

According to Zosi Automotive Research, in 2025 integrated cockpit-driving cars will reach 1.67 million units sold—a 43% increase year-on-year—but only a minority have truly implemented a single-chip solution.

The real significance of this calculation lies in the mid- and lower-end market. The 100,000–200,000 yuan range dominates China’s auto market, and suburban NOA (Navigation on Autopilot) remains a rare feature in this segment—not because the algorithms aren’t capable, but because the cost of having two domain controllers is unsustainable. The BOM saved by integrated cockpit-driving solutions is enough for automakers to make advanced smart driving a standard, not optional, feature. The Wen Dao V9 is a case in point: an MPV starting at 194,800 yuan, with city NOA and full-scenario automatic parking as standard across the range, features that two years ago only appeared in 300,000-yuan models. Whoever helps automakers achieve this first wins the fixed-point contracts.

This track is getting increasingly crowded. In the Qualcomm 8775 camp, Desay SV’s integrated domain controller has brought the cost of smart cockpit plus smart driving below 6,000 yuan, and has secured contracts with Chery and Tata. Bosch is simultaneously adapting to both Qualcomm and MediaTek. Carlink, Huayang, and ThunderSoft collectively showcased their 8775 solutions at the Beijing Auto Show. Outside the camp, Horizon’s Starry chip uses 5 nm technology with 650 TOPS, debuting with iCAR. Black Sesame’s Wudang C1200 is tied to Dongfeng. Qualcomm’s own SA8797 boosts performance to 320 TOPS, with contracts signed by Li Auto and Leapmotor.

The 8775 solutions aren’t stable yet, but pressure from next-generation chip iterations is already here.

But computing power isn’t the only barrier. At the Beijing Auto Show, some integrated cockpit-driving vehicles already launched suffered from inflated computing specs and laggy user experience. Problems like virtualization isolation, computing power scheduling, and thermal management—these engineering hurdles can only be resolved through repeated mass production. Zhuoyu’s experience across three vehicle types is indeed rare for now.

According to Zhuoyu’s data, by March 2026, it will have 20 customers, covering 32 brands, with over 50 mass-produced models and more than 100 fixed-point breakthroughs. The client list covers half of China’s auto market. Integrated cockpit-driving is only one expansion path for Zhuoyu, but probably the most crucial.

The shelf-life of experience is limited. As more solution providers complete their first mass production and more powerful chips flatten engineering difficulty, first movers' moats will become shallower. Zosi predicts a 36% compound annual growth rate for this market from 2026 to 2030, with plenty of room for increments but limited seats on the track. The mass-production pace in the next year or two will most likely determine how long each supplier stays at the table.

Risk Disclaimer and Exclusion ClauseMarkets are risky, investment requires caution. This article does not constitute personal investment advice and does not take into account individual users' specific investment objectives, financial situations, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their particular situation. Investing based on this is at your own risk.