Don’t forget that energy infrastructure is a recurring cost. The panels only last ~30 years, and other equipment will need refreshing on a similar lifecycle. This alone is a reason that solar energy will never be “dirt” cheap.
(But it might still be a lot cheaper than current fossil and nuclear sources).
Also:
> massive energy-consuming use cases
One such use case is the synthesis of hydrocarbons for plastic/chemical production. We’re going to need a lot of solar to power that.
I've just learned, the term of art for energy cost factoring in labor, infrastructure, replacement, etc is "levelized cost" or LCOE and by that metric utility solar is indeed already cheaper than fossil fuels and nuclear power:
It looks like utility solar is currently ~33% the LCOE price per kwh of nuclear or 85% of the price of gas (CCGT).
This is an increase relative to 2021 (20% of nuclear / 60% of gas)
(As I understand it, the cost increase between 2021 and 2023 is from cost of capital and shipping)
Naively projecting out the LCOE trajectory, it indeed does not look like it's going to get to be 'dirt cheap' soon - eyeballing, the LCOE curve over time appears to be flattening towards a 5% or smaller cost reduction year over year.
> One such use case is the synthesis of hydrocarbons for plastic/chemical production
Why? There are plenty of hydrocarbons left. All climate science and renewable energy policy indicates that we have to stop burning it before we exhaust supplies.
(But it might still be a lot cheaper than current fossil and nuclear sources).
Also:
> massive energy-consuming use cases
One such use case is the synthesis of hydrocarbons for plastic/chemical production. We’re going to need a lot of solar to power that.