Gasgoo Munich- On Nov. 20, at the 7th Next-Generation Powertrain Technology Innovation Summit 2026, Chen Xueping, Chief Engineer of Powertrain Intelligence at the Geely Automotive Research Institute, noted that the boundaries between intelligent driving, cockpit, chassis, and powertrain domains are increasingly blurring. Traditional rule-based control systems can no longer meet the demands of AI-level autonomous decision-making, adaptive optimization, and multi-domain coordination — a gap that serves as the fundamental driver behind Geely's push for powertrain intelligence transformation.

Chen Xueping, Chief Engineer of Powertrain Intelligence at the Geely Automotive Research Institute
On the industry collaboration front, following the 2025 release of the "White Paper on the Preliminary Vision and Exploration of AI Cars" with the China Society of Automotive Engineers (China SAE), Geely partnered with the same organization again in October 2026. At the SAECCE AI Main Forum, they unveiled the "AI Car Powertrain Intelligence White Paper," which outlines a three-stage generational path: 1.0 Control Intelligence, 2.0 System Intelligence, and 3.0 Energy Autonomy. The framework also establishes a three-layer technical architecture centered on "energy efficiency and safety""""comfort and convenience"".
In terms of product deployment, the Xingrui AI Cloud Powertrain has evolved through four stages: "MAP-based traditional control — 2021 Cloud Powertrain 1.0 — 2025 AI Cloud Powertrain 2.0 — multi-agent proactive intelligence"". Version 2.0 is built around perception, decision-making, and execution, establishing dual profiles for the driver and the vehicle. By leveraging scenario engines, electrochemistry, and time-series large models, it achieves multi-objective optimization — shifting from a "one-size-fits-all"" approach to a "personalized for every driver"" experience. The next-generation system will use digital twins and AI surrogate models for global optimization, ultimately realizing "hyper-personalized adaptability"".
Four major function clusters are already in production. Smart energy management delivers roughly 10% fuel savings during long-distance holiday trips. Smart motion control cuts slip by 50% and boosts tracking stability by 15%, with anti-slip intervention engaging within 2 milliseconds. Smart health management extends maintenance intervals by up to 50% and improves battery health by over 15%, while achieving an accuracy rate above 95% for electrified powertrain fault warnings. Smart charging covers home battery maintenance and long-distance charging planning.
On the computing infrastructure side, the Xingrui Intelligent Computing Center 2.0 boasts a combined computing power of 23.5 EFLOPS, making it the largest self-built intelligent computing center among Chinese automakers. Meanwhile, Xingrui PowerMind — a full-domain energy large model designed for charging scenarios — has rolled out three services: AI fast charging, AI healthy charging, and AI automatic charging.
Looking ahead, Chen predicts the powertrain domain will evolve along four trajectories: proactive agent delivery, global optimization via digital twins, cross-domain integration into intelligent entities, and an ecological, full-lifecycle closed loop. As the physical foundation of embodied AI, the powertrain will ultimately transition from a mere driving component into a node connecting users with the broader energy ecosystem — compelling automakers to build a dual competitive edge in "energy + AI"".









