AI AccelerationCo-Processor Chips

The AKD1500 Edge AI Co-Processor accelerates on-device AI deployment with ultra-low-power Akida neuromorphic technology.

800 effective GOPSMaximum throughput
<1 mW/GOPPower efficiency
PCIe · SPIHost interfaces
Download the Product BriefOrder AKD1500
BrainChip AKD1500 Edge AI co-processor chip


Image

Notable Features

  • Designed for Low-Power AI Neural Network Acceleration
    • Unsurpassed images/second/watt
    • Maximum 800 Effective GOPS at <1mW/GOPS
  • On-Chip Learning
    • Personalized Edge AI systems
  • Industry Standard Development Environment
    • TensorFlow/Keras and Pytorch APIs
  • Cost Effective

Two Ways to Integrate theAKD1500 Co-Processor

Configuration 1

AKD1500 with PCIe Interface

Explore PCIe Configuration →AKD1500 block diagram: application processor connected over the PCIe host interface

Configuration 2

AKD1500 with SPI Interface

Explore SPI Configuration →AKD1500 block diagram: microcontroller connected over the SPI host interface

Related Products

Next-Generation AKD1500 PCIe Development Card for Edge AI Development

Integrates BrainChip’s AKD1500 Co-Processor in a standard PCIe form factor for efficient on-device AI inference.

Discover MoreOrder Development Card

Related Products

AKD1500 M.2 Card B+M Key

Add ultra-low-power neuromorphic AI to compact embedded systems with the AKD1500 M.2 Card, designed for easy integration with Raspberry Pi 5 and compatible hosts.

Discover MoreOrder AKD1500 Card

Ready to Integrate the AKD1500 Co-Processor?

The AKD1500 can be paired with any host CPU or MCU via PCIe or low-power SPI interfaces to create very compact, low-power, portable and intelligent devices for applications such as IoT, Industrial, Automotive, and Medical.

Order AKD1500

Applications

Drone flying above the clouds, representing vision and imaging at the edge

Vision & Imaging

Low-power processing supports visual recognition across ADAS, drones, robotics, and surveillance.

Smart speaker in a modern living room, representing consumer and smart home devices

Consumer & Smart Home

Always-listening voice, presence, and appliance intelligence for responsive smart homes; no cloud.

Patient wearing a health sensor, representing personal health devices

Health & Personal Devices

On-device health analysis enables continuous, personalized and adaptive monitoring at ultra-low power.

Industrial motor with monitoring overlay, representing industrial IoT

Industrial IoT

Low-power sensor analysis enables predictive maintenance and environmental monitoring at the edge.

Develop with MetaTFTM

Accelerate, Optimize, and Deploy with Ease

BrainChip’s unique MetaTF software flow enables developers to compile and optimize their chosen models for the AKD1500.

MetaTF development flow: data and models from Edge Impulse, Keras/TensorFlow, ONNX and PyTorch compiled through MetaTF to the Akida model library and runtime for on-chip deployment

  • Convert and optimize your models for Akida deployment
  • Streamline the development pipeline from training to production

Image

Bring Real-Time Edge AI to
Your Next Design

Download the complete AKD1500 Edge AI Co-Processor Datasheet and explore how it can power your next-gen Edge AI design.

Access Datasheet in the Developer Hub