苹果发布会上没说的都在这了：A19 Pro、散热和自研基带

Core Insights - Apple has launched the iPhone 17 series, featuring four new models and the new A19 chip, which is seen as a routine upgrade despite some performance enhancements [1][3] - The A19 chip utilizes a third-generation 3nm process, aligning with prior rumors about its manufacturing [7] - The A19 Pro's CPU performance is reported to improve by approximately 10% compared to the A18 Pro, which is consistent with market expectations [7] - A significant upgrade in the A19 Pro is the introduction of a GPU neural network accelerator, which reportedly increases peak GPU performance to three times that of the A18 Pro [8][10] A19 Chip Features - The A19 chip's GPU neural network accelerator is designed to enhance efficiency by reducing data exchange between the GPU and the Neural Engine, addressing resource wastage [8][9] - The GPU neural network accelerator is intended to improve performance in mixed workflow scenarios, such as AI tasks and gaming [9][10] - The A19 chip also integrates a display driver unit for the standard iPhone 17 model, enabling a 120Hz refresh rate for the first time [17][18] Design Changes - The iPhone 17 Pro has shifted from a titanium frame to an aluminum one to accommodate a new active cooling system, which is a first for iPhones [11][12] - The new vapor chamber cooling system is designed to manage the heat generated by the A19 Pro chip, showing Apple's commitment to maintaining high performance [14][15] - The design changes aim to enhance the device's performance in high-load scenarios, including gaming, image processing, and AI tasks [15] Market Strategy - The standard iPhone 17 has outperformed the Pro model in pre-sales, indicating a successful strategy to enhance its competitiveness through key upgrades [19] - The introduction of the iPhone Air, featuring the A19 Pro chip and new design philosophies, serves as a testing ground for market demand for lightweight devices [20][24] - Apple's ongoing development of in-house chips, such as the C1X modem, aims to reduce reliance on Qualcomm and enhance performance [21][23]