The wires leaving a solar power plant will be built for the maximum power output, and the output would presumably be at or near zero at night. Converting this to nuclear means the power plant can provide maximum output at any time of day. This means more energy can be sold over the year on the same wires.
The maximum output of just about any solar power plan tends to be significantly (~a order of magnitude) less than that of any (current) nuclear plant - hiwhc would mena that upgrades would be needed... (Smaller reactors are closer though)
Fuel cells are a horrible idea for grid scale storage. Just burn the hydrogen in a turbine or something. There's no need to get expensive fuel cells because there's no need to save on weight or volume. On top of that a fuel cell produces DC so there would have to be inverters to use this electricity on the grid. The energy density of hydrogen isn't much of a concern at a power plant, they have room for low pressure tanks.
Fuel cells have significantly better efficiency than burning it and converting motion to electricity. (The inverters would take some of that, but there are large scale solutions from things like HVDC transmission lines)
Nuclear with storage makes a lot of sense. With current, running almost all the time reactors (which does not have storage available),solar can reduce the demand on peaking plants. (I'm mainly thinking PV, since that is the only solar that is relatively cheap)
This is a load curve for my country for yesterday. Around 21GW of nuclear (the current, no-storage type) would make sense, with enough solar / wind / dispatchable peaking (the main current options here is likely diesel and coal) to get to around 30GW or so... (If nuclear with storage, like Terrapower's planned design is used, you could probably install around 24GW or so of nuclear and do quite a bit of the peaking with that (which is currently done with some PV, diesel OCGT, pumped storage and coal (and some wind that seems to peak around the evening peak))