> I complained about the word "unprecedented".
If you want to be a stickler, what that actually means is that it has never happened before. Have we ever installed ~2500 GW of new photovoltaic generation capacity over a twenty year period? Of any individual generation method?
> Have all the factories that got us to this level of (obviously) installed capacity burned down recently? Have the investors who invested in them and the engineers who built them all died?
It's not just a question of supply. It's also a question of demand.
Subsidizing five times as much capacity would require five time as much government money, which may not be available. If the subsidies end, the demand curve takes a hit.
In addition to that, the fact that solar only generates during the day hasn't yet become an issue because the daytime demand is higher, so solar can currently be efficiently used to satisfy the demand differential between day and night. But once there is enough solar generating capacity to satisfy the full differential, that efficiency ends and you start having to deal with expensive energy storage or paying capital and maintenance for alternative generation which is idle and unproductive half the day, effectively requiring the cost of solar to fall lower than the alternative's fuel/operating cost only and not the total cost including capital and maintenance, because the capital and maintenance then have to be paid anyway to make the capacity available at night.
It has also been getting installed more in the locations amenable to it -- places with more sunlight, clearer skies and less expensive land. Notice that a lot of the existing capacity is in places like Arizona and not places like Seattle. But once the demand in those locations is satisfied, you have to move to places without those characteristics and you get less power at higher cost.
Meanwhile just because you have the existing supply doesn't mean you have the existing supply. You need the price to fall to the level that it becomes economical to generate power even for use at night in Seattle, but that price may be lower than the price that customers in Asia and South America would pay, so even if the supply exists, it ends up going somewhere else first.
All this to say that extrapolating from an exponential curve is generally wrong sooner or later, and frequently sooner.
> Solar capacity will rise to 100% of today's total capacity, and it still won't displace lots of other energy sources. Why? Because when power is cheaper and has fewer negative externalities, we'll use more of it.
If it becomes genuinely cheaper even in northern cities and even with the cost of energy storage, it will displace lots of other energy sources and we'll use more of it. But that's still speculation that may or may not ever actually happen. And maybe it does, but it's still justifiable to be prepared if it doesn't.