If application- and transmission-protocol-level DoS vectors are fixed, then you're left with just the raw "lots of traffic" volumetric attacks, which means your attacker has to have a lot of compromised hosts (or the right compromised hosts with lots of bandwidth). I'd say that's a reality that would be easier to handle, because you raised the bar from anyone who can develop or use a script and deploy to a few low-power systems, exploiting protocol shortcomings, to only those who have a bunch of higher-powered systems.
The smaller hosting companies may still very well go out of the game if the problem worsens, even if most DoS venues do end up being mitigated. I don't know how I would respond to that as of this moment, but hopefully it doesn't have to come to that. It's already tough to find a decent hosting company in my experience.