BrainChip

Lore & Background

BrainChip was founded in 2004 by Peter Van Der Made. In March 2015, Australian mining company Aziana acquired BrainChip, and via a reverse merger of the now dormant Aziana in September 2015, BrainChip was listed on the Australian Stock Exchange. Van Der Made then began commercializing his original idea for artificial intelligence processor hardware. In 2016, the company appointed former Exar CEO Louis Di Nardo as CEO, with Van Der Made taking the position of CTO.

The company's MetaTF software is designed to work with image, video, and sensor data for applications including security, surveillance, autonomous vehicles, and industrial automation. It uses Python to create spiking neural networks or convert other neural networks to SNNs for use on the AKD processor hardware, and can also deploy SNNs on normal processors. The AKD1000 processor is an event-based neural processing device with 1.2 million artificial neurons and 10 billion artificial synapses, containing 80 nodes that communicate over a mesh network, each node consisting of four either convolutional or fully connected Neural Processing Units with individual memory units. The processor runs an entire neural network executing all neuron layers in parallel, designed for inference and incremental learning on edge devices with lower power consumption.

In October 2021, BrainChip announced it was taking orders for its Akida AI Processor Development Kits, and in January 2022, for its Akida AI Processor PCIe boards. In April 2022, it partnered with NVISO. In November 2022, it added the Rochester Institute of Technology to its University AI accelerator program. The next month, BrainChip was a part of Intel Foundry Services. In January 2023, Edge Impulse announced support for BrainChip's AKD processor. On 29 January 2023, BrainChip announced completion of the design of its AKD1500 reference chip. On March 6, 2023, it announced the second generation of its Akida platform, adding support for 8-bit weights and activations, a Vision Transformer engine, and hardware support for a Temporal Event-Based Neural Net. On 12 March 2023, it announced that the Akida processor family integrates with the Arm Cortex-M85 processor.

Reader's Guide

BrainChip's significance lies in its development of neuromorphic computing hardware and software that mimics the human brain, specifically through spiking neural networks. The company's AKD1000 processor, with 1.2 million artificial neurons and 10 billion artificial synapses, uses event-based processing to analyze essential inputs at specific points, storing results in on-chip memory. This design allows inference and incremental learning on edge devices with lower power consumption, a key advantage for applications like security, surveillance, autonomous vehicles, and industrial automation. The MetaTF software enables training and deployment of SNNs, including conversion from other neural networks, and can run on normal processors as well as the AKD hardware.

BrainChip's legacy is shaped by its partnerships and milestones: joining Intel Foundry Services and the Arm AI partnership, collaborating with NVISO and Edge Impulse, and adding the Rochester Institute of Technology to its university accelerator program. The company's second-generation Akida platform, announced in March 2023, added support for 8-bit weights and activations, a Vision Transformer engine, and hardware support for a Temporal Event-Based Neural Net, while the AKD1500 reference chip design was completed in January 2023. These developments position BrainChip as a participant in the edge AI and neuromorphic computing space, with a focus on low-power, event-driven processing.

Did You Know?

• BrainChip was founded in 2004 by Peter Van Der Made and listed on the Australian Stock Exchange in September 2015 via a reverse merger with mining company Aziana.
• The AKD1000 processor contains 1.2 million artificial neurons and 10 billion artificial synapses, with 80 nodes communicating over a mesh network.
• BrainChip is a member of Intel Foundry Services and the Arm AI partnership.