JDelekto
Regular
The Edge boxes are like "mini-PCs" that provide compute services in a small form factor, usually using processors specifically designed for lower power consumption, like those in laptop PCs. These boxes, however, are specialized to perform AI functionality. Mini-PCs have been used in corporate conference rooms running camera and voice conferences, as well as those running home entertainment systems.This links to a Qualcomm site, just to give an idea what the use cases for an Edgebox are:
Edge AI Box
Our Edge AI computing solutions help accelerate the deployment of new applications across IoT use cases: security and surveillance, retail, smart factories, smart cities and more.www.qualcomm.com
Seems to be more than just another Devkit and as such could have an financial impact.
Hard to say if this qualifies for an Ann, but at first sight it looks like price sensitive.
The Raspberry Pi BrainChip offered was provided as a platform to evaluate, develop, and test Akida's technology. While hardware hackers and hobbyists could use it for their projects, it was not intended as a consumer retail item. The offering through VVDN is a product for a specific market, and one such example could be for home automation (like controlling lights, alarms, and smart-cooling systems.) They could also be used in security centers monitoring video feeds, but there are many other uses.
I think it's important to note that this is a vehicle to sell a product in a relevant market using BrainChip's IP while BrainChip is still in the business of selling commercial IP. I've recently read the Motley Fool article where they speculate that the market is spooked by Intel's announcement of their new "Meteor Lake" processor that will contain an NPU allowing users to run apps like Stable Diffusion and ChatGPT (most likely using the Llama engine). This can technically be done now with PCs running a decent Nvidia desktop or mobile processor.
Intel has had a "Compute Stick" product, a USB device that uses their Movidius processing VPU, deep learning hardware designed for processing video. Additionally, others have used this processor in a PCIe form factor. Intel has probably now leveraged some of its neural network hardware as a module in its processors, making it an "all-in-one" package specifically designed for mobile and Edge devices (like Edge boxes).
Let's not forget BrainChip's business model and mantra: licensing IP and making sensors smart. BrainChip's goal is to provide AI processing for event-driven sensors that process different types of input on a broad spectrum. This is an entirely different market than Edge PCs or laptop processors. Smart sensors, like our digital smartphones, have a much more prolific base, and just as Qualcomm penetrated several industries with its low-power processors, I believe that BrainChip can do the same with its neuromorphic capabilities.
That their technology can be used in an Edge box and compete with other giants like Nvidia and Intel only shows its adaptability from its intended target market.