What solar and wind have made irrelevant is the assumption that constant supply of the baseload is an addressable market for a power plant. That only works if the supply is the cheapest on the market (something nuclear has even historically struggled to achieve without big subsidies), and solar and wind will supply the baseload much cheaper than any other source when they're running, causing a lot of problems for any plant on the grid that wants to run at a constant output.

Data centers are only 5-10% of all US electric power demand. A significant fraction of that load is responsive to spot market price. DCs incorporate their own battery storage infrastructure for backup purposes and some even sell stored energy back to the grid during times of high demand.

it is not at all a perfect match for nuclear. Solar produces incredibly cheap power for the daytime hours, pushing the nuclear off the grid (why would a data center buy expensive nuke power in the hours when cheap renewables are available?). Nuclear needs to be running and selling power at its high price 24/7 for it to be even close to economically viable. Only supplying power that people want(due to its high price) during the night is fatal to nuclear.

Thats what OP means by baseload generation being irrelevant.

Why should your data center buy multiples more expensive nuclear powered electricity when renewables or stored renewables are available? They won’t.

And even if they sign a PPA the renewable arbitrage they could do exerts large downward pressure on the price they are willing to pay.

Which is to say, they would want to sign a contract saying ”deliver cheap reliable electricity when renewables and storage doesn’t do it”.

And now you are trying to fit an extremely CAPEX heavy square into a round OPEX sized firming hole.