Investment Rating - The report does not provide specific investment recommendations for individual companies or suppliers in the intelligent driving sector [4][9]. Core Insights - 2025 is identified as a pivotal year for automotive intelligence, with a projected increase in domestic electrification penetration rates to 50%-80% over the next three years, leading to a reshaping of the automotive landscape [4]. - Major intelligent driving manufacturers have successfully implemented complex urban NOA (Navigation on Autopilot) experiences, enhancing high-level functionalities such as parking and scene understanding [4]. - The report evaluates the intelligent driving experiences of ten manufacturers, including XPeng, NIO, and Tesla, through both large sample and small sample road tests, focusing on various dimensions such as scene implementation and comfort [4][9]. Summary by Sections Road Test Overview - The report includes both large sample concentrated road tests and small sample deep road tests to assess the performance of intelligent driving systems [8][27]. - The large sample tests involved approximately 50 participants testing various models along a standardized route, while the small sample tests were conducted by the same evaluators under similar conditions [9][27]. Intelligent Driving Models Tested - The models tested in September 2025 include XPeng P7, NIO ES8, and Tesla Model 3, among others, with specific versions noted for each [10][11]. Performance Evaluation - Compared to Q1 2025, Q3 2025 shows improvements in intelligent driving capabilities across all manufacturers, with a narrowing gap between the leading and following manufacturers [4]. - The report highlights that new entrants in the self-research sector are showing promising performance, with significant iterations expected in the coming months [4]. Specific Model Insights - XPeng's XOS 5.7.7 demonstrated excellent performance with an average takeover frequency of 1.51 times, showcasing strong capabilities in complex scenarios [34]. - The NIO ES8 with cedar model achieved an average takeover frequency of 2.03, indicating robust performance in urban environments [43]. - Tesla's FSD version 13.2 recorded a higher average takeover frequency of 5.73, suggesting areas for improvement in its intelligent driving capabilities [49]. Technical Developments - The report notes advancements in hardware and software across various manufacturers, with many now utilizing self-developed chips and algorithms [26][24]. - Specific improvements in driving comfort and efficiency were reported, with some models achieving significant enhancements in user experience metrics [22].
9月上海篇:2025年主流车企城市NOA试驾报告