Re: Not really a surprise
"various estimates reckon we'll need around 3x more electricity than we generate now."
And many other estimates put it alot lower.
Your surge pricing is an absurd illustration of what won't happen - grid scale storage is not limited to chemical batteries, though they are clearly the best option for instant and short term grid balancing.
The cost of storage and of demand shaping is already visible on the grid, because we have tariffs which track wholesale rates and we have systems like DFS.
You're welcome to come here and have a cup of tea.
Dunkelflaute is an issue, but it's not insurmountable, not by a long stretch - for some places in the world it simply doesn't exist, and for others it's something we're dealing with. Of course the solutions haven't been built yet, but they will be. And a combination of different storage technologies (and this is the main situation where I can see hydrogen having a place) will be used to deal with differing duration events. One of those technologies may well include demand management - easy enough for my energy supplier to top up my car before the event and then delay recharging until after, there will obviously be people who need to charge during the event, but we can pretty easily shift load away.
Heating is usually the critical load here - and in the short term we might have to burn some gas, but we get more heat into people's homes by burning that gas at a power plant than we would by burning it in a boiler. At the point at which you can shape demand extensively... then you can shape it to be flat, or to any other shape you want.
The issue is that nuclear isn't particularly dispatchable - it's continuous, but it doesn't scale up and down with demand easily. Now smaller reactors have an advantage here, but they're still far better off running at their design power constantly. That constant generation does increase the lowest generation during a DF, but it only slows down the rate at which we'll consume storage, and probably not by a huge amount. If we had 25% nuclear then we'd drain storage at a slightly slower rate, but we'd also be less able to fill it (since we'd have less "other" generation). If we can get to micro-reactors, then I think we're potentially onto a great solution for data centres - which have an effectively constant demand (though the hyperscale providers can relatively easily move load to a region with lower electricity prices).