If it was just galactic rotation curves that would be explained by adding a (mostly) non-interacting dark matter particle, then MOND would probably have many more proponents. It's harder to understand how to replace non-interacting particles in explaining the cosmic-microwave-background power spectrum, galaxy clusters, and structure formation. Maybe it's possible to make a complicated theory do those things, but at that point, I'm not sure it wins on simplicity anymore.