Theoretically yes... It requires two main things:
- new code generator
- new objects with alternative code implementations eg not relying on arm neon
Neither really hard, but also not trivial.
I do wonder about the viability, as the code is open source, it be difficult to sell - so how to fund the development work.
Note: Johaness is the only one who could make a commercial version ( say pro version) from the existing code base without making the derived code own source - this a bit limits the 'inventive' for a third party/independent developer doing it.
Id question that, I think the reality might be in drives a few sales, but this this would be nothing compared to the number of users using it to create VSTs, and therefore the support would be relative to these numbers... arguably this would skew the focus considerably. .... not saying thats bad or good, really more a matter of if this is a direction @johannes wants to go, rather than a technical challenge.
(of course, all just my opinion, using the knowledge I have of axolot/vst dev)