AI@Edge Compact Fanless Embedded BOX PC with Google Edge TPU


  • GbE PoE/PD x 1
  • USB3.2 Gen 1 x 2 & USB2.0 x 2
  • Mic-in & Line-out
  • Support MEMS x 2 (Optional)
  • Built-in WiFi 2x2 MIMO (802.11b/g/n/ac 2.4/5GHz) & Bluetooth 4.1
Visión general

Awarded the Taiwan Excellence Award for 2021, the BOXER-8521AI is focused on providing flexibility in deployment and connectivity. The BOXER-8521AI features a PoE PD port, allowing the system to be deployed further away from its power source, as well as enable internet connection and remote monitoring of the system over the same single cable, reducing the complexity of installation. Additionally, by utilizing both the PoE PD port and DC-input, the system can continue operating even if one power supply is cut off.

The BOXER-8521AI powers AI Edge Computing thanks to the Google Edge TPU System-on-Module (SoM) with a rugged chassis and flexible I/O loadout. The Google Edge TPU combines the NXP i.MX 8M SoC (quad Cortex-A53, Cortex-M4F) with the Google Edge TPU coprocessor to deliver speeds up to 4 TOPS at an energy efficient 2 TOPS per Watt.

Thanks to the Google Edge TPU, developers and users who deploy the BOXER-8521AI have access to Google’s suite of training software, AutoML Vision Edge. This innovative software helps developers quickly develop and train AI models utilizing Google’s vast cloud services, and then export the model to the BOXER-8521AI and Google Edge TPU. The system supports TensorFlow Lite, a framework which allows for more efficient models, better system performance and processing speeds.

The BOXER-8521AI is built rugged, with fanless construction to keep dust and contaminants out, providing longer-lasting, reliable operations. Additionally, the system can operate in temperatures from -5°C to 50°C without loss in performance. The BOXER-8521AI is equipped with a flexible I/O loadout that includes COM, HDMI, two USB3.2 Gen 1 and two USB2.0 ports. It also features a 40-pin multi-I/O port to connect to more sensors and controllers, such as cameras, temperature sensors, and others.



AI Accelerator Google Edge TPU ML accelerator coprocessor
CPU NXP i.MX 8M SoC (Quad-core Cortex-A53, plus Cortex-M4F)
System Memory 1GB LPDDR4x
Storage Device 8GB eMMC
Display Interface HDMI 2.0 x 1
Ethernet 10/100/1000Base-TX x 1

HDMI 2.0a x 1

USB Type A x 2 for USB3.2 Gen 1 USB Type A x 2 for USB2.0

RJ-45 x 1 for GbE LAN & PoE/PD(802.3at) 12Vdc (DC Jack w/lockable)

Mic –in x 1 Line-out x 1

40-pin I/O x 1 (GPIO/I2C/SAI/SPI/UART/PWM) DB-9 x 1 for RS-232/485

OS Flash port x 1 (Micro USB)

Micro-SD slot x 1

Indicator Power LED x 1
OS Support ACLinux 4.0 (Compatible with Debian 10)

Power Supply

Power Requirement PoE/PD (802.3at) 12Vdc (DC Jack w/lockable)


Mounting Wallmount
Dimensions (W x D x H) 6.92“ x 3.94“ x 1.54“ (175 .8 mm x 100 .0 mm x 39 .0 mm)
Gross Weight 2.31 lbs. (1.05 kg)
Net Weight 1.28 lbs. (0.58 kg)


Operating Temperature 23°F ~ 122°F (-5°C ~ 50°C, according to IEC60068-2 with 0.5 m/s airflow)
Storage Temperature -40°F ~ 185°F (-40°C ~ 85°C)
Storage Humidity 95% @ 40°C, non-condensing
Anti-Vibration Random, 3.5Grm, 5~500Hz
Anti-Shock 50G peak acceleration (11 msec. duration,eMMC,microSD)
Certification CE/FCC class A
Para detalles del empaque, especificaciones tecnicas y otra informaciónes del producto, descargue la hoja de datos o el manual para el usuario.