Well, I think the main difficulty with SSR is you'd think it would be more of an incremental step. I don't think it's so much a "separation of concerns" issue (i.e. just having some of your presentation logic running on-server doesn't mean it's necessarily not isolated logic) as it is a "wow I have to rewrite significant parts of my application to take advantage of this" issue. Some frameworks definitely do a better job with this than others IMO, but not everybody wants to (or should) migrate their React app to NextJS (for example) just to get SSR.