From what I can make out of this is that You’re knees deep into overthinking your setup.
Fundamentally, You shouldn’t be pursuing the 1000 brilliant new community ideas you have all at once by spinning up as many new discourse (and by extension any other software) instances.
community building is an iterative process and needs focus, if you split your focus between more than one community at a time, You will end up with more on your plate than you can handle.
My suggestions were a food for thought, as I mentioned earlier when I did these experiments these were just that, a bunch of friends doing things because we can.
there is a reason (or two) why these are not common practices? These come with a significant maintenance and management overhead that will cause a lot of headaches and sleepless nights down the line, you have been warned.
given your objectives, I would rest my case with this statement: what you’re trying to achieve is not impossible, but today is not the day to achieve everything that’s possible. Focus on what you have at hand, build your first community, the rest can follow on when the time comes.
As for learning by experimenting, if cutting costs is your primary goal, multisite is the way to go, with all the compromises it presents.