does mean that you shouldn’t give him a pass

Understood, and agreed. Not giving him a pass at all.

Energy generation: The sunlight isn’t changing nor is the solar installation’s size.

I do not accept the premise that the “solar installation’s size” is not changing. Quite the contrary, I am calling for it it increase. A lot. (All of this applies to wind as well, but for simplicity, I’m going to focus on solar alone. We need more wind, and more wave, tidal, etc.) One of the factors I am trying to address is suboptimal generation conditions: overcast skies and seasonal variations reducing expected output. Overbuilding solar production relative to demand means that normal demand can still be met under less than ideal conditions.

you’d still need the factories to be willing to take on the increased energy costs associated with moving into the more expensive energy window.

I do not accept the premise that the new window is more expensive. We are already regularly experiencing negative rates in these windows because demand is not high enough to match existing production, let alone the increased production I am calling for. These negative rates are extraordinarily bad for continued solar development; they cannot be allowed to continue. That means we either slow/stop expanding solar, or we drive existing and/or new demand to daylight hours.

Slowing solar rollout is not at all an option, so we are left with some variety of demand shaping.

It is important to note that the underlying reason rates are actually going negative is because baseload generators can’t roll back their output to match daily demand curves. They have to stay online, and at high output levels to be able to meet (artificially inflated) overnight demand. If they tried to roll back, they would not be able to roll forward fast enough. Removing that excess overnight demand would allow them to lower their total output. The least efficient baseload generators are coal-fired; with less overnight demand, some of these remaining coal-fired plants can be retired.

Making the change you and OP are referring to works in a bubble but not in the way you are envisioning it.

The missing component is elastic demand. Bitcoin has some serious negative connotations, and I am not recommending this as a specific example of what we should be adding. Rather, I ask you consider only the nature of the load: they are turning electricity into money, and they can only do that profitably at certain price points.

Consider massive data centers on highly variable rate plans. They fall well below the rates that regular consumers would pay when there is a surplus. But, they also rise far above market rates when there is a shortage. A Bitcoin miner operating on such a plan would shut down when their instantaneous costs rise above their expected returns.

Consider a data center with its own solar facility supplementing the power they draw from the grid. What will this facility do when the instantaneous price of power exceeds their rate of return? Will the continue to draw power at a loss? Will they reduce their demand and continue operating solely on their own solar generation? Or, will they shut down their miners entirely, and use their solar to backfeed the grid at that high price point?

Another possible flexible load is desalination: don’t try to store the power; use the power when it is available, and store the desalinated water instead.

Hydrogen electrolysis is another option. It’s not particularly efficient, but when you can get the power at a heavily discounted rate, the effective economic efficiency may be high enough to justify it. Don’t try to store the power; store the hydrogen instead.

Fischer-Tropsch hydrocarbon synfuel production is yet another option: don’t try to store the power; store the generated fuel.

AI is another (controversial) option. Perform the energy-intensive training operations when power is cheap, and suspend processing when rates rise.

All of these operations may need to be under the control of the grid provider, to provide sufficient incentive for them to be shut down when production falls unexpectedly.