Core Insights - Advantech is showcasing its integration capabilities with NVIDIA Jetson Thor at COMPUTEX 2025, focusing on intelligent mobility, Vision Language Model (VLM) at the edge, and humanoid robotics [1][6] - The MIC-743 Edge AI platform is highlighted for its compact, rugged design and optimized performance for edge applications [1][5] Edge AI Architecture - The system integrates a full sensor-to-inference pipeline, supporting various camera interfaces and utilizing NVIDIA Holoscan Sensor Bridge for low latency applications [2][3] - Data is processed through Advantech's high-speed sensor switch and streamed into the MIC-742, which is built on the NVIDIA Jetson Thor module [3][4] MIC-743 Features - The MIC-743 is designed for harsh environments, offering higher AI performance, lower power consumption, and a smaller footprint compared to traditional x86 systems [5][9] - It supports a range of applications from heavy-duty machinery to humanoid robotics, enabling real-time performance for vision-based decision-making [6][9] Ecosystem Integration - Advantech's experience in edge architecture allows for a ready-to-deploy Edge AI solution, reducing system complexity and deployment time [6][9] - The integration of AI hardware, sensor infrastructure, and networking facilitates multi-tasking capabilities [6][9] Product Specifications - The NVIDIA Jetson Thor Module features 2,070 FP4 TFLOPS AI performance, 128GB LPDDR5X memory, and various I/O options for high-bandwidth sensor input [8] - The rugged design of the MIC-743 allows for operation in temperatures from -10°C to 60°C and withstands vibrations [8]