Nvidia(US:NVDA) 美股IPO·2025-08-26 00:31Core Viewpoint - Nvidia's Jetson Thor platform significantly enhances AI computing power and efficiency, enabling real-time processing for robotics and physical AI applications [1][3][5]. Group 1: Product Features and Specifications - Jetson Thor offers up to 2070 FP4 TFLOPS of AI computing power with a power consumption of 130 watts, and features 128GB of memory [1][3]. - The platform's AI computing performance is 7.5 times greater than its predecessor, Jetson Orin, while CPU performance has increased by 3.1 times [5][7]. - The developer kit, Jetson AGX Thor, is priced at $3499, with a bulk purchase price of $2999 for orders over 1000 units [1][3]. Group 2: Applications and Industry Adoption - Jetson Thor is designed to support various generative AI models, enabling real-time processing of high-speed sensor data and visual reasoning in dynamic environments [7][12]. - Major industry players such as Agility Robotics, Amazon Robotics, and Boston Dynamics have adopted Jetson Thor for their next-generation robotics [3][12]. - The platform is expected to enhance the agility, decision-making speed, and autonomy of robots, making it crucial for navigation and interaction in real-world scenarios [13]. Group 3: Ecosystem and Support - Jetson Thor is supported by a complete Nvidia Jetson software platform, compatible with major AI frameworks and generative AI models [11]. - The platform has attracted over 2 million developers and more than 150 hardware and software partners since its launch in 2014 [11]. - Research institutions like Stanford University and Carnegie Mellon University are leveraging Jetson Thor to advance capabilities in perception, planning, and navigation models [11]. Group 4: Autonomous Driving Applications - Nvidia has also launched the DRIVE AGX Thor developer kit for autonomous vehicles, which is now available for pre-order and expected to ship in September [4][14]. - Leading automotive companies, including BYD and Volvo, are developing solutions based on the DRIVE AGX Thor platform [14][15].