By "cheap" you also left out the author's assumption that "the propulsion system will require a tiny fraction of the total gained mass to convert into energy, which we assume could not be a problem for a Type-II civilization."

So they already start with a Type II civilisation that basically dismantled part of their planetary system to reach that state in order to... build probes that construct more Dyson Swarms? Still doesn't sound like a very compelling argument to me.

The paper also doesn't deal with the actual expansion aspect - only with the detectability of micro probes moving through- and replicating within star forming regions, i.e. not places you'd actually want to colonise - neither inside or outside of gravity wells.

> Subspecies and cultural divergence can be seen as a solution, leading to more innovation that can be shared at the speed of light.

You seem to greatly underestimate the vastness of space. Even a tiny (by Fermi-Paradox standards) interstellar realm would have worlds hundreds of lightyears apart. There's not much use in sharing "innovations" if they arrive after 500 years - not much innovation left after such a long time (unless you assume that one side stops innovating entirely for some weird reason, yet keeps listening for signals), plus it'd become increasingly harder to even keep a common cultural understanding given the temporal and spatial distances. Language changes over time, for example.