{"product_id":"dx-m1-ai-accelerator-m-2-module-with-4gb-lpddr5-25-tops-df-dfr1252","title":"DX-M1 AI Accelerator M.2 Module with 4GB LPDDR5 (25 TOPS)","description":"\u003cp\u003eThe DX-M1 AI Accelerator M.2 Module delivers server-class edge computing power, packing 25 TOPS (INT8) of inference performance into a standard \u003cb\u003eM.2 2280\u003c\/b\u003e form factor with ultra-low power consumption (\u003cb\u003e2W–5W\u003c\/b\u003e). Equipped with \u003cb\u003e4GB LPDDR5\u003c\/b\u003e memory and a PCIe Gen3 x4 interface, this NPU ensures low-latency processing and seamless compatibility with \u003ca href=\"https:\/\/www.dfrobot.com\/kit-003.html\" target=\"_blank\"\u003eRaspberry Pi 5\u003c\/a\u003e, \u003ca href=\"https:\/\/www.dfrobot.com\/topic-278.html\" target=\"_blank\"\u003eLattePanda\u003c\/a\u003e, and various x86\/ARM platforms. Supported by the comprehensive \u003cb\u003eDXNN® SDK\u003c\/b\u003e for PyTorch, ONNX, and TensorFlow, it is the ideal solution for deploying complex AI models in intelligent robotics, visual SLAM, and industrial automation systems.\u003c\/p\u003e\u003cp\u003e\u003c\/p\u003e\u003cp style=\"text-align: center; \"\u003e\u003cimg src=\"https:\/\/dfimg.dfrobot.com\/enshop\/20251118\/1.jpg\" alt=\"DX-M1 AI Accelerator M.2 Module Functional Block Diagram\" title=\"DX-M1 AI Accelerator M.2 Module Functional Block Diagram\"\u003e\u003c\/p\u003e\u003cp style=\"text-align: center;\"\u003eFigure: DX-M1 AI Accelerator M.2 Module Functional Block Diagram\u003c\/p\u003e\u003cp\u003e\u003cb\u003e\u003cbr\u003e\u003c\/b\u003e\u003c\/p\u003e\u003cp\u003e\u003cb\u003eServer-Class Inference at the Edge\u003c\/b\u003e\u003c\/p\u003e\u003cp\u003eThe core strength of this M.2 AI module lies in the ability to deliver \u003cb\u003e25 TOPS of INT8 performance\u003c\/b\u003e, a figure previously reserved for power-hungry server components. This massive computational headroom allows for the execution of complex neural networks and multi-stream video analysis directly on the edge device without relying on cloud connectivity. Despite this high performance, the advanced architecture ensures energy efficiency, operating within a strictly controlled 2W-5W power envelope, significantly reducing heat generation and extending operation time in mobile robotic platforms.\u003c\/p\u003e\u003cp\u003e\u003cbr\u003e\u003c\/p\u003e\u003cp\u003e\u003cb\u003eUniversal M.2 Compatibility \u0026amp; Integration\u003c\/b\u003e\u003c\/p\u003e\u003cp\u003eDesigned with the industry-standard \u003cb\u003eM.2 M-Key (2280)\u003c\/b\u003e interface, the accelerator card ensures broad interoperability across various computing platforms. It utilizes the PCIe Gen3 x4 protocol (backward compatible with x1 mode) to maximize data throughput. This standardization allows system integrators and developers to easily upgrade existing x86 PCs, LattePanda single-board computers, or Raspberry Pi 5 setups (via HATs), instantly adding extensive AI capabilities to standard hardware infrastructure.\u003c\/p\u003e\u003cp\u003e\u003cbr\u003e\u003c\/p\u003e\u003cp\u003e\u003cb\u003eSeamless Development Toolchain\u003c\/b\u003e\u003c\/p\u003e\u003cp\u003eThe hardware is backed by the robust DXNN® SDK, which streamlines the transition from model training to edge deployment. The toolchain provides a complete environment for compilation, optimization, and runtime execution, removing the traditional barriers associated with NPU development. Popular deep learning frameworks including \u003cb\u003ePyTorch, TensorFlow, TensorFlow Lite, Keras, and XGBoost \u003c\/b\u003eare natively supported, allowing algorithms to be ported via ONNX format with minimal friction.\u003c\/p\u003e\u003cp\u003e\u003cbr\u003e\u003c\/p\u003e\u003cp\u003e\u003cb\u003eIndustrial-Grade Reliability\u003c\/b\u003e\u003c\/p\u003e\u003cp\u003eBeyond raw performance, the DX-M1 is built to withstand rigorous operating environments. The module features \u003cb\u003e4GB of onboard LPDDR5\u003c\/b\u003e memory to handle large models and heavy batch processing efficiently, alongside 1Tbit of QSPI NAND Flash for firmware stability. With an operational temperature range spanning from -25°C to 85°C, the device is qualified for industrial automation, outdoor security monitoring, and other harsh application scenarios where consumer-grade electronics typically fail.\u003c\/p\u003e\u003cp style=\"text-align: center; \"\u003e\u003ciframe frameborder=\"0\" src=\"\/\/www.youtube.com\/embed\/AHjXnjYb4hE\" width=\"640\" height=\"360\" class=\"note-video-clip\"\u003e\u003c\/iframe\u003e\u003cbr\u003e\u003c\/p\u003e\u003cp\u003e\u003c\/p\u003e\u003cp style=\"text-align: center;\"\u003eVideo: Butter Benchmark: AI Chip Battle - 94% Energy Cost Savings Proven!\u003c\/p\u003e\u003cp style=\"text-align: center;\"\u003e\u003ciframe frameborder=\"0\" src=\"\/\/www.youtube.com\/embed\/Js6Soex0WI4\" width=\"640\" height=\"360\" class=\"note-video-clip\"\u003e\u003c\/iframe\u003e\u003c\/p\u003e\u003cp style=\"text-align: center;\"\u003eVideo: \u003ca href=\"https:\/\/www.youtube.com\/@deepx2692\/videos\" target=\"_blank\"\u003e[DEEPX] Introduction to DXNN (DEEPX SDK)\u003c\/a\u003e\u003c\/p\u003e\u003ch2\u003eApplications:\u003c\/h2\u003e\u003cli\u003eRobotics: Visual SLAM \u0026amp; Navigation\u003c\/li\u003e\u003cli\u003eEdge Computing: Real-time Video Analytics \u0026amp; Object Detection\u003c\/li\u003e\u003cli\u003eIndustrial: AI Visual Inspection \u0026amp; Safety Monitoring\u003c\/li\u003e\u003cli\u003eAutonomous Systems: Drone Perception \u0026amp; Self-Driving Platforms\u003c\/li\u003e\u003ch2\u003eSpecification:\u003c\/h2\u003e\u003cli\u003eProcessor Performance: 25 TOPS (INT8)\u003c\/li\u003e\u003cli\u003eInterface: M.2 M-Key, PCI Express Gen3 x4 (compatible with x1 mode)\u003c\/li\u003e\u003cli\u003eMemory: 4GB LPDDR5, 1Tbit QSPI NAND Flash\u003c\/li\u003e\u003cli\u003ePower Consumption: 2W ~ 5W\u003c\/li\u003e\u003cli\u003ePower Range: 3.3V±5%\u003c\/li\u003e\u003cli\u003eFramework Support: PyTorch, ONNX, TensorFlow, TensorFlow Lite, Keras, XGBoost\u003c\/li\u003e\u003cli\u003eOperating Systems: Windows 10\/11, Ubuntu 20.04\/22.04 LTS\u003c\/li\u003e\u003cli\u003eOperating Temperature: -25°C ~ 85°C (Throttling); 25°C ~ 65°C (Non_ Throttling)\u003c\/li\u003e\u003cli\u003eProduct Dimensions: 22 mm x 80 mm x 4.1 mm\/0.87 inch x 3.15 inch x 0.16 inch\u003c\/li\u003e\u003ch2\u003eDocuments:\u003c\/h2\u003e\u003cli\u003e\u003ca href=\"https:\/\/wiki.dfrobot.com\/SKU_DFR1252_DX-M1%20AI%20Accelerator?heisgoodman\" target=\"_blank\"\u003eProduct WIKI\u003c\/a\u003e\u003c\/li\u003e\u003cli\u003e\u003ca href=\"https:\/\/deepx.ai\/wp-content\/uploads\/2025\/09\/17172223\/2025-0917-DEEPX-DX-M1-M.2-LPDDR5x2-AI-Accelerator-E-Brochure.pdf\" target=\"_blank\"\u003eProducts Brief\u003c\/a\u003e\u003c\/li\u003e\u003cli\u003e\u003ca href=\"https:\/\/docs.radxa.com\/en\/aicore\/dx-m1\/dx-sdk\/dx-model-zoo\" target=\"_blank\"\u003eModel Zoo V2.0.0\u003c\/a\u003e\u003c\/li\u003e\u003cli\u003e\u003ca href=\"https:\/\/docs.radxa.com\/en\/aicore\/dx-m1\/quickly-start\" target=\"_blank\"\u003eQuick Start Guide\u003c\/a\u003e\u003c\/li\u003e\u003ch2\u003eProduct Includes:\u003c\/h2\u003e\u003cli\u003eDX-M1 AI Accelerator M.2 Module with 4GB LPDDR5 x1\u003c\/li\u003e","brand":"DFRobot","offers":[{"title":"Default Title","offer_id":42343285030994,"sku":"DF-DFR1252","price":317.71,"currency_code":"AUD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0735\/0383\/files\/DFR1252_Main_01.jpg?v=1774161974","url":"https:\/\/www.pakronics.com.au\/products\/dx-m1-ai-accelerator-m-2-module-with-4gb-lpddr5-25-tops-df-dfr1252","provider":"Pakronics®","version":"1.0","type":"link"}