I think this is a major (the only?) reason these internet companies are even successful. Because they just ignore all these costs that don’t scale well that ordinary companies should have.
I think that is a touch strong. Though, I agree that largely nobody really factors in costs of scale, as absurd scale just isn't something that makes sense. This is true in physics as much as it is in economics. Things that work in one scale do not in another. Things that work in changing scale don't work in fixed scale.
This is what I would immediately suspect to start. Probably a good start, even. Does run into complications of jurisdiction and hazards of growth. None reasons not to do something.
I'm sympathetic. I think I even agree. Policy will be tough to craft here, sadly.