How can engineers test neuromorphic AI on compact devices? A new developer board pairs an ultra-low-power accelerator with familiar embedded interfaces.
BrainChip and Neuromorphyx have introduced the BrainBoard1500, a compact developer and evaluation board built around BrainChip’s AKD1500 neural-network accelerator. The board was designed to help developers evaluate spiking neural network (SNN) workloads in embedded systems where power, latency, size, and direct sensor connectivity are important.
The board follows the Arduino Nicla form factor and provides direct connector compatibility with the platform. An open-source Arduino library and drivers provide a familiar software path for developers working with embedded sensors and microcontrollers.
At the hardware level, the board provides a direct SPI/QSPI connection for running SNN models on sensor data, while I2C is used for system control. Onboard flash provides storage for models, allowing inference workloads to be evaluated without relying on a larger host computer.
A two-channel current monitor is also integrated into the board. This allows developers to measure the power consumed by AKD1500 inference workloads in real time and evaluate different sleep and wake configurations. Such measurements can help engineers understand how neuromorphic processing behaves under the power constraints of battery-operated edge systems.
The standard SPI/QSPI architecture also allows the board to be connected to a range of embedded hardware ecosystems. The companies said the interface could support software adaptations for platforms from Raspberry Pi, Seeed Studio, OpenMV, Espressif, DFRobot, Adafruit, SparkFun, and STMicroelectronics, along with custom microcontroller and FPGA systems.
The initial software release includes the Neuromorphyx Arduino library, drivers, and Akida Engine integration. BrainChip also plans to support model repositories targeting applications such as keyword spotting, visual wake words, and human activity recognition.
Neuromorphyx developed the board architecture, electronics, PCB design, and Arduino-oriented software, while BrainChip supplied the AKD1500 and technical support. “BrainBoard1500 is exactly the kind of partner-led product that extends Akida’s reach into developer communities,” says Sean Hehir, CEO, BrainChip.
Click here for the official announcement.
