GSA 2026 | AutoAI Unveils Full Product Matrix

 Gasgoo Munich- The "2026 Shanghai International Low Carbon Smart Mobility Exhibition" (GSA 2026) kicked off on July 2 at the National Exhibition and Convention Center in Shanghai. AutoAI took the stage with a full product matrix — ranging from smart cockpit domain controllers to cabin-driving fusion units, and from AI Boxes to central computing platforms. The showcase highlights future-ready system-level solutions, aiming to drive a green, intelligent mobility revolution through AI.

Image source: AutoAI

On the domain control front, AutoAI showcased several heavyweights: the AL-A2 cabin-driving fusion controller, the AL-A1 cabin-driving fusion controller, the AL-C2 premium smart cockpit controller, and the AI Box.

The AL-A2 cabin-driving fusion controller is built for next-generation central computing architectures. Powered by the Snapdragon 8797 platform, it offers enhanced computing power and AI expansion capabilities. The unit supports unified processing for smart cockpits and advanced driver-assistance systems (ADAS), alongside on-device large model deployment and multi-domain coordination. It is designed to handle diverse scenarios, including high-level autonomous driving, Robotaxis, and embodied intelligence.

The AL-A1 cabin-driving fusion controller stands as the world's first mass-produced single-chip solution of its kind. Built on the Snapdragon 8775 platform, it handles both cockpit and intelligent driving tasks on a single System-on-Chip (SoC) while integrating a 4.4-billion parameter on-device AI model. The solution has already entered mass production in models like the BAIC Arcfox Alpha T5, the S5, and the Arcfox Wendao V9.

Based on the Snapdragon 8255, the AL-C2 premium smart cockpit controller focuses on delivering comprehensive digital services for OEMs. Paired with the Autosee smart cockpit, it fuses in-cabin and external perception data to support multi-screen linkage, 3D interaction, and immersive experiences. The system is already in mass production in vehicles such as the Jetour Zongheng G700.

Notably, AutoAI’s AI Box — an on-device AI computing unit — can integrate independently with existing cockpit platforms. By packing capabilities like large model inference and multimodal perception into automotive-grade standalone hardware, it offers a low-cost upgrade path for legacy vehicles. This allows automakers to meet the AI upgrade demands of existing users without requiring extensive modifications to the original vehicle systems.

Additionally, AutoAI highlighted the AI Link 3.0 Deep Fusion EEA central computing architecture, which made its global debut at CES earlier this year. It is the industry’s first electronic-electrical (E/E) architecture to achieve deep fusion at the system level, marking a shift from "multi-domain collaboration" to a new stage of "full-domain deep fusion."

Comprising a central computing platform, zone controllers, and a high-speed optical communication backbone, the Deep Fusion EEA boasts system-level performance metrics including 256 Gbps of bidirectional bandwidth and end-to-end latency of less than 1 microsecond. This enables efficient coordination of perception, decision-making, control, and interaction, laying a systematic technical foundation for future scenarios such as high-level autonomous driving, on-device large models, and in-vehicle AI Agents.