The issue is that nuclear scales well with size, but that increases iteration time & costs for a single iteration. It also leads to much less efficiency possible compared to e.g. solar cells. So once they didn't improve and were still expensive (and getting more expensive, like all big construction projects) the potential seems limited. Same as start ups getting huge evaluations while running deficits, but no-one giving the benefit of the doubt at that scale to a failing 30-year old business with promising basic ideas.
Every (big honkin') nuclear power plant is more-or-less a giant project onto itself. The hope of nuclear advocates is that with small reactors, their reduced cost & increased scale (in terms of units) will more than make up for their small size (and resulting loss of scaling in that dimension). I think that's unlikely, but some startups are trying so we'll see if it's a winning formula.
My argument is: Why bother? Renewables are cheap, storage technologies[1] have a lot of viable ways of going forward and are getting better very quickly. Even if through a lot of investment nuclear energy gets (economically, build-time-wise) viable in 20 years like renewables are now, I doubt it would change much or solve pressing issues.
[1]: Starting from common chemical batteries, over to hydrogen/power-to-gas, molten-sand and a variety of concepts exploiting potential height energy. The limited commercial adoption yet can be attributed to the fact that they're not necessary yet, so the added price is too high. Germany, with a very high share of renewables and planned to be growing, is only starting to dip it's toes with some very basic "stabilizer" like in Australia and nothing big-scale in the near future. If this was a pressing issue, you'd think governments would do more (Especially in germany, the cost structure strongly disincentives storage right now, needing to pay certain costs twice, while sinking and sourcing the energy).