The problem is that solar and wind requires backup generation, usually CCGT natural gas.
This is ok until you scale higher and higher. You end up having CCGT only producing 20-40% of the time (to fill in for wind and solar blips). This massively increases the capex of CCGT plants, as you're only producing rarely. This would get more and more extreme with more solar+wind penetration.
Another massive problem is solar+wind overproduction, which is really hard to solve and is starting to really hit the German, UK and California grids. On very sunny AND windy days you get massive energy overproduction. You then have to either:
a) Turn off solar+wind remotely (often very expensive to retrofit to existing installations as the Germans found out) b) Pay other (natgas, coal, nuclear, etc) operators to shut down, which can be very expensive c) Hope that negative electricity prices make more demand. This is unlikely to happen as industrial users can't switch on extra production quick enough to respond to this.
This is made worse by feed in tariffs being paid at any energy price, incentivizing solar+wind to continue generating even if electricity price is negative (say it is -€0.05kWh spot, but your FIT is €0.20/kWh, you are still going to produce as you will net 0.15euro per kWh.
The two 'solutions' which are often mentioned are battery storage and HVDC long distance transmission.
Battery storage is still horrendously expensive on a kWh basis. It may come down, but this is an enormous problem. I am personally not sure there is enough lithium left that is easily extracted to make this viable at the scale (billions of kWh) required.
HVDC connections I also am suspicious of - if it's sunny and windy in Germany, it is likely to also be the same 1000km away more or less.
If we're talking about problems that show up when wind/solar becomes an order of magnitude more prevalent than it is now, why would you care about the cost of retrofitting anything?
Even in your entirely hypothetical example you say they're having to pay natgas and coal to stop production.
Is it just me or is that really weird?
I mean turning off coal and gas is kind of the point, isn't it?
If you could turn gas and coal up and down in a few minutes, then perhaps you would have a point. But coal can take 6-24hrs and natgas 1-4hrs. It won't suddenly stop when you have a massive gust of wind.
More realistic excuses I've heard are long term contracts, potentially with minimum run times for fossil plants. Which are therefore the actual problem at the moment.
The lion's share of the costs are upfront and the break even period is measured in decades. What is the energy landscape going to look like in 30 years? That's a question that you need to be able to answer with high confidence if you want to correctly price bonds that will be used to construct a nuclear power plant to be paid off from the revenue of the plant. If the potential bond buyers can't answer that question with high confidence then they don't know what interest rate is appropriate for the bonds and they probably won't want to buy them.
Grid-scale storage is one possibility, but more work needs to be done to make it viable. For technologies that exist today, nuclear looks the best for base load capacity.
While natural gas is better than coal, it's still not good from an emissions perspective compares to nuclear.