Data Input Interfaces

  • PCI Express 2.1 x2 Lane Endpoint
  • USB 3.0 Endpoint
  • I3S, I2C, UART, JTAG

On-Chip Processor

  • M-Class CPU with FPU & DSP
  • System Management
  • Akida Configuration

Data Processing

  • Pixel-Event Converter
  • SW Data-Event Encoder
  • Any multivariable digital data
  • Sound, pressure, temp., others

External Memory Interfaces

  • SPI FLASH for boot/storage
  • LPDDR4 Program/Weights

Multi-Chip Expansion

  • PCIe 2.1 2 lane root complex
  • Connects up to 64 devices

Flexible Akida Neuron Fabric

  • Implements 80 NPUs
  • All Digital logic with SRAM (8MB)
  • Also Available as Licensed IP Core
  • First Implementation: TSMC 28nm

Akida AKD1000 is a reference chip implemented with TSMC at 28nm

Contact Sales

Reference Chip
as a Hardware Accelerator

Place an AKD1000 with its accompanying software development ecosystem next to an MCU or CPU where it can accelerate AI compute tasks.

BrainChip’s software development ecosystem and runtime are seamlessly compatible with any CPU hardware and are OS agnostic.

MCU / CPU

AKD1000
Hardware Accelerator

Akida AKD1000 has several key features differentiating it from other neural network processors and deep learning accelerators.

  • Event-based computing leveraging inherent data and activation sparsity
  • Fully configurable neural processing cores, supporting convolutional, separable-convolutional, pooling, and fully connected layers
  • Incremental learning after off-line training
  • On-chip few-shot training
  • Configurable number of NPUs
  • Programmable data to event converter
  • Configurable on-chip SRAM memory
  • Runs full neural networks in hardware
  • On-chip communication via mesh network
  • On-chip learning in event domain
  • Process technology-independent platform
  • Network size customizable to application needs
  • IP deliverables include: RTL, dev tools, test suites, and documentation
  • Event-based NPU engines running on a single clock

AI Enablement
Program

Learn more

BrainChip’s AI Enablement Program makes entry to edge AI simple and real.
Each of our tiered programs brings you from concept to working prototype with varying levels of model complexity and sensor integration. In addition, our AI experts provide training and support to make the process efficient and smart.

Learn more