Unboxing the bundle
BrainChip has taken its AKD1500 M.2 neuromorphic accelerator and wrapped it in a ready-to-run Raspberry Pi 5 bundle, as CNX Software reports. The new AKD1500 Raspberry Pi 5 Dev Kit ships as a complete system in a metal case, pairing a Raspberry Pi 5 8GB, an official M.2 HAT+, and an Akida AKD1500 M.2 B+M-Key card. A 32GB SD card comes preloaded with Ubuntu Server 20.04 LTS (marked TBC by CNX Software), Python 3.7, and the MetaTF development environment. Most Pi connectivity is preserved: the Gigabit Ethernet port, four USB ports, two HDMI outputs, and USB-C power connector remain accessible through the case, along with the Pi's power button and an LED on the M.2 HAT+.
One notable choice is the camera requirement. Rather than depending on a MIPI CSI ribbon-cable camera, the kit expects a USB camera, which simplifies setup for many users even if it means one extra cable.
Demos and local inference
Once the board is connected to Ethernet or Wi-Fi, users open the kit's IP address in a web browser to reach the Akida Demo Dev Kit launcher. All demos run locally on the device, with no cloud round trip.
The included demos are:
- Visual Wake-Word (VWW) person presence detection
- YOLO object detection
- MobileNet edge demo
- MobileNet face demo
For custom work, BrainChip's MetaTF ML framework is the main path, and the product brief mentions Edge Impulse support, though CNX Software notes that Edge Impulse does not appear in the user manual. A more interesting hardware-level feature is on-chip training: the AKD1500 can learn new classes directly on the device without cloud retraining, which fits the low-power, low-latency neuromorphic pitch.
The price premium for convenience
The turnkey experience is attractive, but the price is hard to ignore: $1,295. CNX Software did the component math and found a DIY equivalent could be assembled for roughly $350 before taxes and shipping.
- Raspberry Pi 5 8GB — about $175
- M.2 HAT+ — about $12
- Akida AKD1500 M.2 card — $129
- Metal case (e.g., Geekworm P579) — about $10.99
- 32GB microSD card — $10–$20
- Estimated assembly and software setup time: around two hours
That means the kit is not cost-optimized for hobbyists. Instead, it is a support and convenience bundle for engineering teams, educators, and labs that want to evaluate event-based neuromorphic inference on a familiar Raspberry Pi 5 platform without sourcing components or wrestling with software installation.