ArmSoM RK3588 AI Module7 (AIM7) is a Rockchip RK3588 system-on-module compatible with NVIDIA Jetson Nano and other 260-pin SO-DIMM Jetson modules, so you should be able to use it with a wide range of Jetson carrier boards. Â It is designed for edge computing, AI inference, VR/AR, smart NVR systems, and general-purpose applications.
The RK3588 AIM7 ships with 4GB to 32GB RAM, 32GB to 128GB eMMC flash, and integrates a Rockchip RK806-1 PMIC, but no Gigabit Ethernet controller as found on the competing Mixtile Core 3588E. While the RK3588 SoM should work on most Jetson carrier boards, ArmSoM has also designed the AIM-IO carrier board to quickly get started with the module.
RK3588 AI Module7 specifications:
- SoC – Rockchip RK3588
- CPU – Octa-core processor with 4x Arm Cortex-A76 cores @ up to 2.4 GHz, 4x Arm Cortex-A55 cores @ 1.8 GHz
- GPU – Arm Mali-G610 MP4 GPU with support for OpenGL ES3.2, OpenCL 2.2, Vulkan1.1
- AI accelerator – 6 TOPS NPU with support for INT4/INT8/INT16
- ISP – 48MP ISP with HDR support
- VPU
- Video Decoding – 8Kp60 H.265/VP9/AVS2, 8Kp30 H.264), 4Kp60 VP9, 1080p60 MPEG-2/1, VC1, VP8
- Video Encoding – 8Kp30 H.264/H.265
- Process – 8nm
- System Memory – 4GB, 8GB (default), 16GB, or 32GB 64-bit LPDDR4x
- Storage – 32GB (default), 64GB, or 128GB eMMC 5.1 flash
- 260-pin SO-DIMM edge connector compatible with Jetson Nano SoM
- Storage – SDIO 3.0, SATA
- Display – DP 1.2a, HDMI 2.1/eDP, 2x 2-lane MIPI DSI
- Camera – 2x 4-lane+2x 2-lane or 5×2-lane MIPI CSI interfaces @ 2.5Gbps/lane; up to 12x MIPI CSI cameras
- Audio – 2x I2S
- Networking – Gigabit Ethernet
- USB – 1x USB 3.0 Gen1, 3x USB 2.0
- PCIe – PCIe Gen3 x4 + PCIe Gen2 x1
- Low speed I/Os – 3x UART, 2x SPI, 4x I2C, GPIOs
- Power Supply
- Supply Voltage – 5 VDC
- RK806-1 PMIC + discrete power circuit
- Dimensions – 69.6 x 45 mm (12-layer PCB; NVIDIA Jetson TX2 NX, Orin Nano,… form factor)
- Weight – 17.3 grams
- Temperature range – Operating: 0°C to +80°C; storage: -20°C to +85°C
- Humidity – Operating: 10 to 90%; storage: 5 to 95%
AIM-IO carrier board specifications:
- Supported SoM – RK3588 AIM7 SoM and other Jetson Nano/NX compatible modules
- Storage – MicroSD card slot
- Video Output
- DisplayPort and HDMI ports
- 4-lane MIPI DSI up to 4K @ 60 Hz
- Camera
- 2x 2-lane MIPI CSI connectors
- 1x 4-lane MIPI CSI connector
- Up to 8x camera supported
- Networking – Gigabit Ethernet RJ45 ports
- USB
- 4x USB 3.0 Type-A ports
- USB Type-C port for firmware flashing
- Expansion
- 40-pin GPIO header
- M.2 E-key (PCIe/USB/SDIO/UART) socket for WiFi or other modules
- Power Supply
- 12V DC via power barrel jack
- 5V PoE
- Dimensions – 100 x 80 x 29 mm (Same footprint as Jetson Nano developer kit, compatible with enclosure)
- Weight – 78.4 grams
ArmSoM provides Debian 11 and Ubuntu 24.10 images for the RK3588 AIM7 SoM and AIM-IO carrier board, and highlights support for the RKNN-Toolkit2, converting TensorFlow/PyTorch models to RKNN format to leverage the 6 TOPS NPU. The system can perform object detection with YOLOv5s (1080p) at 55 FPS, semantic segmentation with DeepLabv3+ (4K) at 25 FPS,
and video analytics of eight 1080p streams processed synchronously with less than 100 ms latency. You’ll find more details on the documentation website, as well as on Google Drive.
You can check our review of the Mixtile Core 3588E Jetson SoM to get an idea about software support and Jetson compatibility of this type of solution. Note this was done with Ubuntu 22.04 in 2023, so software has improved since then, and the issue we had with the older Jetson Nano Developer Kit A02 carrier may or may not happen on the RK3588 AIM7. The important takeaway is that compatibility is not assured for a specific carrier, and this needs to be tested on a case-by-case basis.
ArmSoM decided to launch the RK3588 AI Module7 SoM and AIM-IO carrier board on Crowd Supply with a $4,780 funding target. The module goes for $239 with 8GB RAM and 32GB eMMC flash, and the AIM-IO carrier board for $36. Price included shipping. I’d recommend getting the AIM-IO board even if you already have a Jetson carrier board to have a known-to-work base. For reference, a Mixtile Core 3558E module with 16GB RAM and 128GB eMMC flash goes for $199 (excluding shipping), and the Turing RK1 goes for $169 in 8GB/32GB configuration. Both modules are also compatible with the Jetson Nano/TX2 NX ecosystem, so the ArmSoM model looks expensive in comparison, but I’ve been told it’s because of recent tariffs and the included shipping fee. Rewards from the crowdfunding campaign are expected to ship on January 14, 2026, onwards.
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
