Tesla's FSD Entry into China Shifts Domestic ADAS Focus to Algorithms and Scenario Adaptation

Over the past two years, China's intelligent driving industry has generally used the number of LiDAR units and computing power metrics as the primary benchmarks for assessing autonomous driving capabilities, leading to redundant hardware configurations. Most automakers have pre-installed high-cost hardware intended for Level 3 (L3) autonomous driving; however, due to the slow regulatory approval and deployment of L3 in China, vehicles have remained functionally at Level 2 (L2) for extended periods. As a result, users neither benefit from advanced autonomous features nor avoid bearing the high maintenance costs and depreciation risks associated with hardware such as LiDAR. The entry of Tesla’s Full Self-Driving (FSD) system into the Chinese market has prompted the industry to re-evaluate the technological pathways for intelligent driving. Tesla’s vision-only approach emphasizes algorithmic optimization and adaptation to local road conditions, avoiding the hardware cost burden associated with multi-sensor fusion systems. Influenced by this shift, some automakers have begun scaling back on redundant hardware investments and are instead intensifying algorithm training based on localized data. Moreover, FSD’s rollout has accelerated the standardization of intelligent driving evaluation criteria and spurred improvements in domestic supply chains—particularly in data operations, computing resource allocation, and technical services. The industry’s focus is now shifting from hardware stacking toward iterative algorithm development, broader scenario coverage, and comprehensive cost control.