Using nuclear instead of solar when solar was available would increase the cost of that power by a factor of 3 or more. That would also be insane.
Nuclear just doesn't have a place on a grid with a large renewable supply.
EDIT: You wrote:
"I agree that would be insane, which is why I said to run the nuclear 24/7 and build enough solar for the extra demand during the day."
This is case #2, using nuclear power during the day when you could have used cheap solar. It makes no sense economically at current nuclear costs.
Wind/solar is only that much cheaper if you ignore storage.
Stop ignoring the possibilities of much cheaper nuclear and things look different anyway. South Korea is already 6X cheaper according to the article, and some GenIV designs are likely to do even better.
Is there a POSSIBILITY of much cheaper nuclear? Sure, anything is POSSIBLE. Is it LIKELY? Experience says "no". And surely it's not possible in the US right now. Who is going to suddenly build these much cheaper nukes here, the Nuclear Power Plant Fairy? The SMR companies are not in a position to deliver now; the one that is closest, NuScale, is not promising the kind of cost reductions that would be needed.
https://www.forbes.com/sites/jeffmcmahon/2018/10/01/3-reason...
Maybe the U.S. will keep dragging its feet, but that doesn't mean the technology is inherently expensive, it just means we're bad at it in the U.S.
(And you're still ignoring storage savings in the solar/nuke scenario.)
There actually is a simple solution: follow UAE and let Koreans build standardized APR-1400.
https://spectrum.ieee.org/energy/nuclear/the-united-arab-emi...
At the same time, solar PPAs in the US are falling so fast it's hard to keep track of the record holder. Here's one from last year for $0.023/kWh.
https://www.greentechmedia.com/articles/read/nevada-beat-ari...
But if we're going to build a lot of wind, that still doesn't necessarily doom nuclear, if we can lower the capital cost of nuclear. The article describes some fairly simple ways to go about that; a recent study at MIT had the same recommendations.
Down the road a decade or so, molten salt reactors will start to be available, and those look to have much lower capital costs and good load-balancing.