Gasgoo Munich-China's intelligent driving sector is undergoing intense consolidation. Resources are concentrating among leading players, and technical paths are converging. However, the market structure remains unstable. Automakers are competing fiercely on production, algorithm iteration, and scenario adaptation. Currently, local startups and domestic brands lead by leveraging full-stack in-house development. They show agility in local adaptation. Most foreign automakers lag due to slower algorithm updates and insufficient localization.
To reverse this trend, Volkswagen Group China is prioritizing its "In China, for China" tech strategy. It is deepening AI collaboration with Horizon Robotics via CARIZON, a local joint venture. Moving away from adapting overseas solutions, this partnership uses a white-box model. It aims to build a proprietary, controllable AI driving system. By advancing mass-production iteration and R&D for Level 3 and Level 4 autonomous driving, Volkswagen has defined its roadmap.
Unlike short-term alliances common in the industry, this partnership is core to Volkswagen closing its intelligent driving gap. It aims to stabilize market share in China. This also marks a significant step for foreign automakers integrating into the local AI ecosystem.
Abandoning the Pure Outsourcing Model
Historically, most foreign automakers relied on off-the-shelf solutions adapted post-purchase. Core algorithms and AI foundations were dependent on overseas headquarters. This approach failed to manage complex local scenarios like city streets and mixed traffic. Consequently, driving experiences lagged behind domestic rivals. Volkswagen’s previous systems faced similar issues: conservative algorithms and slow iteration speeds.
To address this, Volkswagen’s partnership with Horizon Robotics innovates the standard model. It moves beyond a simple buyer-seller relationship. Using CARIZON as a local hub, Volkswagen accesses Horizon’s foundational AI models via white-box authorization. This allows independent development of the full algorithm chain. It covers sensor perception, decision-making, and vehicle control, securing control over technology iteration.
This ecosystem integrates CARIZON’s C7H chips with the GAIA data platform. It creates a closed-loop R&D system unifying hardware, data, and algorithms. It addresses the localization issues of generic overseas solutions. This significantly boosts recognition accuracy and predictive capabilities in complex conditions.
Source: Volkswagen Group
Leadership comments confirm this upgrade's strategic priority. Oliver Blume, Chairman of Volkswagen Group, stated China is the core engine for global innovation. This is especially true for AI and autonomous driving. Localized development is not just for China. Mature technologies will be exported to global markets. Volkswagen now views China not just as a sales market, but as a central R&D hub for intelligent technology.
Compared to peers, Volkswagen emphasizes balancing practicality with safety. All self-developed algorithms adhere to global safety standards. The system meets mass-production needs while reserving interfaces for high-end scenarios like Robotaxis. This lays a foundation for long-term development.
Dual-Track Delivery: Production and R&D
While many automakers struggle with execution, Volkswagen’s upgrade has clear milestones. By pursuing a dual-track approach of iterating current models and reserving future technology, execution efficiency is high.
Regarding mass production, Volkswagen's local technology has entered delivery. CARIZON’s advanced driving assistance system is finalized. It is scheduled to launch in the third quarter of 2026. It will debut on seven new electric models across three joint ventures. The system enables full-scenario navigation assistance. This addresses weaknesses in urban driving capabilities.
Source: Volkswagen Group
Starting in 2027, the system will fully adapt to the CEA. This enables standard upgrades across new energy models, increasing competitiveness. Previously, CARIZON’s system was integrated into the Yu Zhong 06 and 07 models. It rolled out features like highway navigation and memory parking. This accumulated local road data and production experience.
Regarding high-level R&D, collaboration will accelerate Level 3 and Level 4 rollout. Supported by Horizon’s computing power, Volkswagen sped up its timeline. It plans official delivery of Level 3 autonomous driving in the second half of 2027.
Ralf Brandstätter, Chairman of Volkswagen Group China, stated the brand will focus on safety. It aims to build a benchmark high-level intelligent driving system. Relying on local AI models, the company targets rapid and safe deployment.
From a competitive standpoint, this layout addresses Volkswagen’s shortcomings. The 200,000 RMB pure electric vehicle market is fiercely competitive. Domestic brands capture share through intelligent driving advantages. Foreign brands' pricing power continues to erode. By combining local full-stack development with chassis tuning and safety, Volkswagen gains a competitive edge. This combines traditional engineering with local tech. It challenges the perception that foreign automakers lag in intelligence.
