Good blog post by Paul Sztorc doing some calculations / economic modeling on viability of L2s under future transaction demands.
I am too dumb / have not read closely enough to react in more detail, but PS makes his assumptions explicit and walks you through the logic. He also provides some links to previous discussions on the topic that I wasn't aware of, so that will be almost certainly be good context.
If you're into detailed game-theoretic analyses of how the economics of this stuff works, have a peek.