Grove AI HAT: A powerful and compact Raspberry Pi HAT for edge computing workloads that costs less than US$25

Edge computing is nothing new, with its origins dating back to the Advanced Research Projects Agency Network (ARPANET), which the United States Department of Defense established in 1969. The rise of IoT and the sheer volume of data now produced globally outstrips the bandwidth and performance that cloud computing currently offers, so it remains more efficient to compute some things close to the end-user. You may have even experienced edge computing, with Amazon, Apple and Tesla leading the way with their connected services like Alexa, FaceID and Autopilot.

Now, Seeed Studio has made it easier to use the Raspberry Pi with edge computing workloads thanks to its Grove AI HAT, which has four Raspberry Pi HAT mounting holes. The Grove AI HAT can be powered by a Vin pin on a Pi or via its 5V USB Type-C port. The HAT measures just 83 x 55 mm but Seeed Studio has managed to integrate the following:

• Sipeed MAix M1 AI RISC-V module (RISC-V: 400-600 MHz)
• Kendryte K210 dual-core CPU 64-Bit
• 16-bit 230 GMULps Neural Network Processor (KPU)
• Float Point Unit (FPU) that supports DP & SP
• Audio processor (APU) that supports up to 8 microphones
• 1x USB 2.0 Type-C - power & programming
• 1x LCD interface
• 1x Digital Mic
• 1x Accelerometer sensor
• 1x KTAG & ISP UART pin header
• 2x 20 pin header with I2C, UART, SPI, I2S, PWM, GPIO support
• 1x Power LED
• 1x Boot LED
• 1x Boot button
• 1x Reset button

Seeed Studio has added full support for ArduinoCore-API, which it has based on the Kendryte standalone SDK so that you can use the Grove AI HAT with Arduino IDE, Linux, Mac OS X and Windows among others. This allows the HAT to utilise Grove Arduino Libraries and other Arduino libraries too. The Grove AI HAT is currently available to pre-order for US$24.50, with Seeed Studio estimating that it will ship worldwide from June 15. The company has stated that prices will rise after that date to US$28.90.