Surprisingly hard if the confidence intervals are wide enough.

They're not? I mean, combined cycle plants have been built out the wazoo, so there's good data on that (replacing natural gas with hydrogen causes only minor changes, mostly in the combustor, and there are already industrial combustion turbines burning fuel gases that are mostly hydrogen.) Cost data on recent nuclear builds in the west are also widely available.

I suppose you may be meaning there's no good upper bound on how bad the expected unexpected cost overruns in nuclear can be.

I mean I don't have any confidence your numbers are correct.

The cost of combustion turbine power plants is well established since so many have been built. The cost of nuclear plants built recently in the west is also readily available.

Given that, it gives me pause that you don't actually know that information, given how eager you are to represent your "knowledge" on this subject.

The costs of creating 100 of those hydrogen storage caverns is completely unknown. You also have to add the cost of power lines to distribute the power to the rest of the country.

What? Creation of storage caverns is certainly well costed, because that's a standard way natural gas is stored. There are 36 such storage caverns for natural gas in the US. The estimated per energy capacity cost of creating such caverns is as little as $1/kWh (when storing hydrogen).

Hydrogen could also be stored in aquifers and depleted natural gas fields, just as natural gas is stored in those.

What will the leakage rate be? Hydrogen leaks through everything.

Hydrogen has already been stored in underground caverns, so it's clearly acceptable.

Geological storage like this involves very thick layers (hundreds or thousands of feet), so the leakage will be low in any case. If geological formations can trap methane for millions of years, they can trap hydrogen for much shorter times.