This is a Raspberry Pi 5 with active cooler + M. 2 HAT containing an Akida 1000 neuromorphic card. The Akida uses Spiking Neural Networks (SNNs) to simulate the way our brain operates as opposed to using Graphics Processing Units (GPUs) for the implementation of (convolution) neural networks.
Additionally, squeezed inside of the case is a solar charge module which is obviously attached to the 3 18650 batteries on top and the large solar panel.
As to whether this will actually work is anyone’s guess. You never know unless you try.
Power will definitely be an issue as a the Pi 5 expects around 25W. In theory the Pi may run at lower power albeit at reduced performance levels. But, add the Akida and fan and it’s really pushing the limits of available power. Although the Akida is supposed to use a minimal slice of the available power – a key advantage of SNNs – reputedly 1W?
As ever, research and theory can only get you so far, the proof of the pudding is building real ‘things’, experimenting and really trying your best!
With data centres degrading our planet through intense use of natural resources of all kinds who knows? This could be the future! 😁