Specifications
Board
BrainBoard1500 Standalone AKD1500 Co-Processor Board
Accelerator
BrainChip AKD1500 neural-network accelerator (Akida neuromorphic engine)
Architecture
32-core spiking neural network (SNN) accelerator
Effective performance
Up to 800 effective GOPS
Local memory
1 MB local transfer memory
Host interface
Standard SPI / QSPI — PCIe and serial-capable device class
Form factor
Arduino Nicla form factor
Software (out of the box)
Neuromorphyx Arduino library and drivers; Akida Engine integration
Open-source repository
Arduino software repository, datasheet and supporting documentation included
Compatible hosts
Raspberry Pi, Seeed Studio / XIAO, OpenMV, Espressif, DFRobot, Adafruit, SparkFun, STMicroelectronics
Custom hosts
Open-source adaptation ports to custom MCU and FPGA designs
Compute paradigm
Neuromorphic event-based processing with on-chip learning capability
Power class
Ultra-low-power — designed for compact, battery-constrained edge devices
Evaluation fit
Lets engineers assess the AKD1500 under real power, latency and size constraints of compact edge devices
Differentiator
Unlike larger M.2 or PCIe evaluation hardware, provides direct embedded host integration
Model repository
BrainChip customer-ready models: keyword spotting, visual wake words, human activity recognition
Partner stack
Integrated into Neuromorphyx NeuroReflex stack for event-based sensing, FPGA preprocessing and neuromorphic inference
Manufacture
Designed, manufactured and sold by Neuromorphyx (Nex Novus d.o.o.)
Announced
19 August 2026 — BrainChip and Neuromorphyx partnership
Target applications
Robotics, space, defence, automotive, industrial, education and embedded AI
RFQ lead time
15-day sample lead time; volume pricing on request
Overview
The BrainBoard1500 is a compact developer and evaluation board designed and manufactured by Neuromorphyx around BrainChip's AKD1500 neural-network accelerator. Announced in August 2026 as a BrainChip and Neuromorphyx partnership, it exists to put neuromorphic silicon into the embedded engineering community's hands in a form factor and software ecosystem they already use.
The AKD1500 is a 32-core spiking-neural-network accelerator with 1 MB of local transfer memory and up to 800 effective GOPS — a co-processor built for event-based, ultra-low-power inference rather than conventional dense tensor throughput. The BrainBoard1500 packages it in the Arduino Nicla form factor and exposes it over a standard SPI/QSPI host interface, which is the key design decision: engineers can measure real power, latency and size behaviour inside a compact embedded system instead of on the larger host-centric M.2 and PCIe evaluation hardware normally used for accelerator assessment.
Software support arrives out of the box via the Neuromorphyx Arduino library, drivers and Akida Engine integration, with an open-source repository, datasheet and documentation alongside. Open-source adaptation ports extend host support to Raspberry Pi, Seeed Studio / XIAO, OpenMV, Espressif, DFRobot, Adafruit, SparkFun and STMicroelectronics platforms, plus custom MCU and FPGA designs. BrainChip supplies customer-ready models for keyword spotting, visual wake words and human activity recognition, and the board slots into Neuromorphyx's NeuroReflex stack for event-based sensing and FPGA preprocessing.
Key Benefits
Direct embedded evaluation: assess the AKD1500 under the real power, latency and size constraints of a compact edge device — not a lab bench host. 800 effective GOPS in 32 SNN cores: event-based acceleration for always-on sensing workloads. Arduino Nicla ecosystem: combine the board with existing Nicla hardware and Arduino tooling. SPI/QSPI simplicity: integrate with almost any MCU or FPGA host via open-source ports. Ready-made models: keyword spotting, visual wake words and human activity recognition ship from BrainChip. Neuromorphic differentiation: on-chip learning and ultra-low power suit always-on, battery-constrained applications.
Applications
Always-on keyword spotting and voice wake-word detection; visual wake words and smart-camera triggering; human activity recognition for wearables and industrial safety; robotics with event-based vision sensing; space and defence embedded AI; automotive in-cabin sensing; ultra-low-power industrial monitoring; education and research into spiking neural networks.
Request a Quote — NEUROMORPHYX BRAINBOARD1500 — BRAINCHIP AKD1500 NEUROMORPHIC DEVELOPER BOARD
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →