> In particular, simple cycle gas turbine power plants with efficiency of 40% cost maybe $400/kW. Compare this to $8-10K/kW for a new nuclear power plant.

The problem being that it only operates ~2% of the time compared to ~100%, and has a shorter operating lifetime in practice, and that isn't counting the cost of storing the hydrogen nor the energy cost to produce it.

Yes, the power will be expensive during that 2%. But it will not contribute all that much to the total cost of operating the grid. In particular, it would be cheaper than forcing the consumers to pay for nuclear the other 98% of the time, just so it would be available during that 2%.

If it's dozens of times more expensive during that 2% because it has to recover 100% of its cost in 2% of the time then it does contribute quite a bit to the total cost of operating the grid, whereas nuclear only has to make up the difference in that time between the market price the rest of the time and its overall average cost.

Storage also has the further disadvantage that you have to over-spec it. It has to be built for the highest capacity you might need and the longest duration, which you don't know ahead of time. If you build less than you need you're in big trouble, but if you build more, you pay for it and get nothing.

Simple cycle gas turbines are 20 times cheaper than nuclear power plants of equal power output. So, no, your argument doesn't hold up when the actual numbers are examined.

Being 20 times cheaper while producing 2% of the kWh because they're turned off 98% of the time makes them 250% more expensive per kWh. And that's comparing the the capital cost for the entire nuclear plant to only the turbine, not including the capital required for the hydrogen storage, or the equipment to produce the hydrogen, or the energy cost of the original generation (with large conversion losses). Or, again, the deadweight losses from the safety margin you need for surplus capacity that you might but probably won't ever use, which could double the cost or more on top of everything else. Whereas if you spec additional nuclear it means you're generating that much more useful electricity 100% of the time.

Meanwhile that kind of storage will have more difficulty finding investors, because if it turns out that some cheaper or better alternative comes along, an investment in nuclear might have to average generating power below levelized cost, but at least you recover most of the capital. Putting in $100 in capital only to have the net present value fall to $80 sucks, but not nearly as much as putting in $100 in capital for a complete write off because you were expecting to be selling to the grid 2% of the time when it turns out to be 0% between demand based pricing and better than expected competing storage technologies. Which means higher capital costs (meaning interest rates) that reduce relative competitiveness even further.

> Being 20 times cheaper while producing 2% of the kWh because they're turned off 98% of the time makes them 250% more expensive per kWh.

Which ignores the cost of using nuclear during that other 98% (or whatever) of the time, when it is very expensive compared to the alternatives.

Yes, the cost just during that 2% is high. But the total cost is lower than if nuclear has to be forced down the consumer's throats 24/7.