China's Smart Driving Goes Global Amid Triple Challenges: Compliance, Data, and Scenario Adaptation

From January to May 2026, China exported 4.059 million vehicles, a year-on-year increase of 63%. Of this total, exports of new energy passenger cars reached 1.792 million units, surging by 120% compared to the same period last year. In the full year of 2025, China’s total vehicle exports amounted to 7.098 million units, ranking first globally for the third consecutive year. As export volumes grow, domestic adoption of intelligent driving technologies is accelerating rapidly: in 2025, the penetration rate of new vehicles equipped with L2+ or higher-level ADAS features stood at 28%, rising to over 41% by April 2026. However, exporting intelligent driving systems faces multiple real-world hurdles. The foremost challenge lies in overseas regulatory compliance. To access the European Union market, automakers must comply with a suite of regulations under the United Nations Economic Commission for Europe (UNECE) framework, including UN Regulation No. 79 (steering systems), ELKS (Emergency Lane Keeping System), UN DCAS (Driver Control Assistance Systems), and UN R157 (Automated Lane Keeping Systems, applicable to Level 3 autonomy). Additionally, the EU’s General Safety Regulation (GSR) imposes stringent validation requirements for features such as Intelligent Speed Assistance (ISA), necessitating extensive local road testing. Data compliance presents a second major barrier. Under regulations like the GDPR, data used to train large AI models for autonomous driving cannot be transferred across borders. Automakers must therefore collect, process, and train models using data gathered locally within target markets. Some countries further require explicit individual consent prior to data collection—a requirement that poses significant practical difficulties. Consequently, certain automakers have opted to initially export vehicles without activating advanced intelligent driving functions, deploying them incrementally at a later stage. A third challenge stems from the algorithms’ limited adaptability to overseas road environments. Significant differences exist between domestic and international conditions regarding parking space configurations, lane marking standards, traffic signage, construction zone setups, and driver behavior. For instance, European lanes are narrower; ceramic stud markers are commonly used in the Middle East; Germany’s unrestricted autobahns demand more robust control algorithms; and shorter following distances in Eastern Europe frequently trigger false alerts. Moreover, some countries enforce unique traffic rules—such as Germany’s requirement to pull slightly off the road when temporarily stopping to allow emergency vehicles to pass—which are difficult to cover through training on domestic data alone. Currently, Chinese automakers are adopting divergent strategies for overseas expansion: 1) **“Sell first, upgrade later”**: Prioritize market entry while deferring deployment of advanced intelligent driving features; 2) **“Comprehensive localization, full functionality from day one”**: Emphasize localized development and user experience; 3) **“Strategic partnerships”**: Collaborate with overseas suppliers to co-build data centers or adopt federated learning and layered map solutions to reduce compliance and development costs. On the policy front, in February 2026, China’s Ministry of Industry and Information Technology (MIIT) and seven other government departments jointly issued the *Guidelines on Automotive Data Cross-Border Transfer Security (2026 Edition)* to standardize outbound data management. Nevertheless, regulatory bottlenecks abroad persist. Industry consensus holds that successful global expansion of intelligent driving cannot rely solely on a “go abroad first, figure it out later” approach. Sustainable operations require simultaneous resolution of three core issues: regulatory compliance and market access, data localization, and adaptation to local driving scenarios.