Vibrating a piston at low frequencies isn't hard, but getting it to move enough air is very hard. You're right that "big enough" with "enough power" could solve that problem, but the mechanics of a conventional driver result in unavoidable output roll-off as you go lower, whereas the fan/rotary solution won't roll off because it has essentially infinite excursion - it just keeps blowing in the required direction at the required pressure level for any low-enough frequency, even fractions of Hz.

Re: custom electronics - my understanding is that the electronics just need to handle very low frequencies (in part by being able to sustain enough current without fading), which would also be true for a conventional subwoofer system trying to operate at those frequencies.

I remember reading about this thing a long time ago and wondering how close I could get with a DIY solution. It would probably sound terrible but it would be fascinating as an transducer learning platform.