AAEON has introduced the BOXER-8741AI, a small but strong edge AI SBC, featuring the new NVIDIA Jetson T5000 module. This device is designed for tough edge AI tasks and is aimed at fields like robotics, self-driving cars, and smart city projects that need lots of computing power and various connection options.
Hardware Overview:
At its centre, the BOXER-8741AI has a 14-core Arm Neoverse-V3AE 64-bit processor and 128GB of fast LPDDR5X memory. This makes it great for AI tasks and handling data. The NVIDIA Jetson T5000 AI accelerator greatly enhances its ability for deep learning projects, and there’s a slot for M.2 2280 M-Key to add fast NVMe storage.
The BOXER-8741AI also excels in connectivity. It includes a QSFP28 port that can handle up to four 25GbE connections, one 5GbE port, and three extra 1GbE RJ-45 ports. This setup is perfect for environments with a lot of sensors and works well with LiDAR, cameras, and industrial tools. It has options for CANBus, RS-485, and GPIO, plus several USB 3.2 Gen 2 connections to plug in different gadgets.
For expanding the system, there are slots for M.2 3052 B-Key, M.2 2230 E-Key, and M.2 2280 M-Key. This allows developers to add wireless components or more storage. For screens, it offers HDMI 1.4a and eDP outputs, plus a CSI-2 header for connecting cameras.
The BOXER-8741AI is made to work in tough industrial settings with a temperature range from -20°C to 60°C (when there’s airflow). It also features built-in platform security and the Aspeed AST2600 Baseboard Management Controller (BMC), which allows for remote management through a web interface.
When we talk about how well the software fits together, this system is compatible with NVIDIA JetPack 7.0. This allows programmers to use the latest AI programs, CUDA libraries, and tools for deep learning.
Price and Availability:
At the time of writing, AAEON’s latest edge AI SBC, the BOXER-8741AI, is in its sample testing phase as of September 2025, with mass production scheduled for November 2025. While AAEON has confirmed the development timeline, they have not yet released any pricing details, a common approach for advanced, pre-release hardware. You can look forward to seeing a sample soon, which we will use for testing or adding to prototypes. We should have news about when we will sell it to customers and how much it will cost by the end of 2025. If you want, I can help track the price changes and let you know when it becomes available for sale. More information is available on the product page.
Sayantan Nandy