I actually know the answer to the technical question, having helped design large scale solar generation in the multi-megawatt range.

The answer is that:

1. Solar panels must be geographically far apart; that way scattered cloud cover won't affect all your panels at the same time; and

2. Build much more solar generation than you need, then turn a bunch of it off almost all of the time. Then as load fluctuates turn on (deploy) solar generation to match the load. As the load reduces turn off (dispense with) generation.

3. Build large scale batteries (which is what GP was alluding to) to charge; then run the demands of the grid from the batteries. Batteries are consumable parts and staggeringly expensive at grid scale. Any large scale use of batteries will literally double the generation costs of PV solar per watt/hour

All of these things dramatically increase the cost per watt/hour of solar generation. Economies of scale are achieved by building large solar installations. If you build them in the desert then transmission lines (something like between $100k to $1m per km) must be constructed. If you build many smaller generation sites where people live, then they are expensive as you lose economies of scale for construction.

Building capacity specifically not to use is the single most expensive use of solar generation. Investors hate it. They want to get returns for every panel they nail onto a frame.

Grid control issues are difficult to solve and are very expensive. Designing a grid that doesn't use the Baseload / Spinning reserve control model is complex and requires rebuilding the biggest and most expensive system ever built.

Facts are not equal to feelings.

I actually know the answer to the technical question, having helped design large scale solar generation in the multi-megawatt range.

The answer is that:

1. Solar panels must be geographically far apart; that way scattered cloud cover won't affect all your panels at the same time; and

2. Build much more solar generation than you need, then turn a bunch of it off almost all of the time. Then as load fluctuates turn on solar generation to match the load.

3. Build large scale batteries (which is what GP was alluding to) to charge; then run the demands of the grid from the batteries. Batteries are consumable parts and staggeringly expensive at grid scale. Any large scale use of batteries will literally double the generation costs of PV solar per watt/hour

All of these things dramatically increase the cost per watt/hour of solar generation. Economies of scale are achieved by building large solar installations. If you build them in the desert then transmission lines (something like between $100k to $1m per km) must be constructed. If you build many smaller generation sites where people live, then they are expensive as you lose economies of scale for construction.

Building capacity specifically not to use is the single most expensive use of solar generation. Investors hate it. They want to get returns for every panel they nail onto a frame.

Grid control issues are difficult to solve and are very expensive. Designing a grid that doesn't use the Baseload / Spinning reserve control model is complex and requires rebuilding the biggest and most expensive system ever built.

Facts are not equal to feelings.