ECARX Zenith Computing Platform Debuts With Qualcomm SA8797P for AI Cockpit and ADAS

ECARX has launched the ECARX Zenith computing platform, marking a significant step in advancing high-performance automotive computing solutions. Built on the Qualcomm SA8797P 4nm chipset, the platform is designed to meet the growing demand for intelligent cockpit systems and advanced driver assistance technologies in next-generation vehicles.

High-performance computing architecture

The ECARX Zenith computing platform delivers a CPU computing power of 560K DMIPS, enabling fast and efficient processing across multiple vehicle domains. This performance level supports complex workloads required for real-time vehicle operations, human–machine interaction, and intelligent decision-making systems.

GPU and NPU capabilities

In addition to CPU performance, the platform offers GPU computing power of 8.1 TFLOPS and NPU sparse computing power of 640 TOPS. These capabilities allow the platform to handle graphics-intensive applications and advanced AI workloads, supporting visually rich cockpit interfaces and AI-driven vehicle functions.

Cloudpeak cross-domain software platform

ECARX’s Cloudpeak cross-domain software platform plays a central role in unlocking the Zenith platform’s heterogeneous computing potential. Its modular design and open middleware architecture enable efficient coordination across computing resources while supporting strong AI performance for automotive applications.

Support for AI cockpit and ADAS innovation

The platform meets automotive-grade safety requirements for instrument panels and ADAS solutions while enabling innovation in generative AI cockpit interaction. It also supports end-to-end ADAS features driven by state-of-the-art VLA models, aligning with the industry’s shift toward AI-centric vehicle architectures.

Compatibility and production roadmap

The ECARX Zenith computing platform is designed for flexible compatibility with global in-vehicle operating systems, including Flyme Auto, Google Automotive Services, and Android Automotive OS. Currently under verification, the platform is scheduled to enter mass production in 2027, positioning it for deployment in future intelligent vehicle programs.