材料汇·2025-07-15 13:31Core Viewpoint - The article emphasizes that the demand for advanced process technology driven by autonomous driving (AD) and embodied intelligence will significantly surpass that of AI GPUs, despite the current hype surrounding AI models like ChatGPT and the performance of companies like NVIDIA [1][2]. Group 1: Wafer Capacity Perspective - The die size of autonomous driving chips is comparable to that of AI GPUs, but the terminal quantity for autonomous driving is several times greater, leading to a much higher demand for advanced process wafer capacity [2][8]. - The value contribution of wafer manufacturing to AI GPU is only 2.25%, indicating that the demand for AI GPUs does not significantly drive wafer capacity needs [10][11]. - The global demand for advanced process capacity for autonomous driving is estimated at 136,200 wafers per month, compared to only 39,700 wafers for AI GPUs [5][6]. Group 2: Application Scenario Perspective - Autonomous driving chips can be viewed as the brain of robots, sharing significant similarities in architecture and application scenarios with robotic intelligence [3][4]. - Companies like Tesla and XPeng are utilizing similar AI chips for both autonomous driving and robotics, indicating a convergence in chip technology across these applications [3][4]. Group 3: Structural Changes in Advanced Process Demand - The anticipated production of robots could reach 1 billion units annually, which, combined with autonomous driving, will disrupt the downstream structure of advanced process applications [4][5]. - The combined demand for advanced process capacity from autonomous driving and embodied intelligence is projected to be approximately 1.65 million wafers per month, significantly exceeding the current capacity of major manufacturers like TSMC [5][6]. Group 4: Die Size and Yield Considerations - The die sizes of autonomous driving chips are generally in the range of 400-600 mm², which is close to that of AI GPUs, but the terminal market for autonomous driving is vastly larger, leading to higher wafer consumption [28][31]. - The yield of larger die sizes is lower, which impacts the overall efficiency of wafer production, making the demand for advanced process capacity even more critical as the industry evolves [39][40]. Group 5: Future Outlook - As the demand for autonomous driving and embodied intelligence grows, the advanced process wafer manufacturing sector is expected to experience a significant expansion, driven by the need for higher performance and more complex chips [6][8]. - The slowdown of Moore's Law suggests that the growth in chip performance will increasingly rely on the volume of chips produced rather than on technological advancements alone [6].