BrainChip's AI for IoT Akida Spiking Neural Network Accelerators Go Mass Market
Designed to mimic the human brain, the Akida accelerator comes with bold claims for efficiency and performance.
Neuromorphic computing specialist BrainChip has announced its biggest milestone yet: full commercialization of its Akida AKD1000 edge AI processors as mini-PCIe accelerators.
"I am excited that people will finally be able to enjoy a world where AI meets the Internet of Things," says Sean Hehir, BrainChip's chief executive. "We have been working on developing our Akida technology for more than a decade and with the full commercial availability of our AKD1000, we are ready to fully execute on our vision. Other technologies are simply not capable of the autonomous, incremental learning at ultra-low power consumption that BrainChip's solutions can provide. Getting these chips into as many hands as possible is how the next generation of AI becomes reality."
The path to commercial availability has been a long one: BrainChip was showcasing its Akida processors, which are designed to mimic the way the human brain operators as a means of running spiking neural networks considerably more efficiently than traditional processors, at the Linley Fall Processor Conference two years ago — but only made development kits, the cheapest of which used a Raspberry Pi as its base and cost nearly $5,000, available to customers late last year.
"We believe the AKD1000 silicon, or the licensing of Akida in a configurable IP format, will lead to major changes in industries using AI at the edge because of its performance, security, low power requirements, and mainly Akida’s ability to perform AI training and learning on the device itself, without dependency on the cloud," claimed Anil Mankar, BrainChip co-founder and chief development officer, at the launch of the development kits.
The commercialized accelerators, thankfully, are considerably cheaper: Pricing starts at $499, with the development kit parts designed to slot in to existing hardware through a mini-PCI Express (mPCIe) slot. The company is also releasing the hardware design files, to make it easier for customers to design their own AKD1000-based accelerator boards — or to integrate the chip directly onto their products as an embedded coprocessor.
The company is now accepting pre-orders for the accelerator boards, on its official website, with delivery expected to begin in around eight weeks' time.