That sort of decoupling could be fascinating. IMO, the walls may be as much a function of economics as physics.

Running a small number of powerful radios on low-frequency spectrum is a cost-effective way to cover large areas. Physics constrains how much of that can happen in the same area. But if networks are willing spend cash, they can get a ton of capacity by densely deploying high-frequency radios. Don't think we're anywhere near hitting the walls on what's possible there.

Putting this another way, networks provide enough bandwidth for everyone in Manhattan. It's just expensive.