Powerful AI Compute
2,000 TOPS on Arm-based architecture, Dual NVDIA DRIVE AGX Thor
Global-Grade Reliability
Certified to ISO 26262, IATF 16949.
Industrial Leading Product
The first L4 auto-grade autonomous driving DCU.
Lenovo L4 Autonomous Driving Domain Controller AD1
Lenovo’s AD1 provides a unified, auto-grade, safety-certified environment for Level 4 autonomy. Built on Arm-based compute and applying Lenovo’s expertise in edge computing and system integration, AD1 delivers a single-chip CPU performance of up to 87 on the SPECrate_int2017 benchmark and AI computing power of 2,000 TOPS. It not only meets the demanding requirements for high-precision perception and decision-making, but also supports flexible computing resource allocation tailored to different scenarios. With its high performance and adaptability, the AD1 excels in complex applications such as robotaxis and unmanned delivery vehicles, providing a solid foundation for the large-scale deployment of autonomous driving.
NVIDIA DRIVE AGX Thor
The NVIDIA DRIVE AGX Thor, powered by Arm architecture, is the most advanced automotive computing platform to date—delivering up to 2,000 TOPS of AI compute performance. Integrated into Lenovo’s AD1 L4 domain controller, Thor consolidates multiple vehicle functions into a single compute unit, including perception, planning, and infotainment. Optimized for generative AI and large-scale vision models, Thor’s Blackwell architecture enhances safety, reduces latency, and accelerates autonomous decision-making in applications such as in WeRide’s Level 4 Robotaxi GXR.
Scaling Level 4 Robotaxis on Arm
Go deeper into how Lenovo and WeRide are moving Level 4 robotaxis from pilot to production using Arm-based compute. Explore the role of HPC 3.0, centralized vehicle architecture, and Arm’s software ecosystem in enabling scalable physical AI.