Good point. Aluminum and copper do not get along well with chlorides, especially when electricity is involved. More troubling is all the talk of one mega-solar farm to rule them all. Centralized power generation is not what the future has in store for many very good reasons. Mega-solar farms do have a role, but not ones covering even half of 525 sq. miles of toxic shoreline.
A key attribute of solar farms is the time of day of production. To make the grid as close to 100% green at the lowest cost, and with the fewest batteries, requires hundreds of sites to buffer the effects of intermittent cloud cover (no matter how rare in the desert) and, more importantly, to protect against natural or manmade disasters. Distributing the solar assets widely provides more hours of active production and protects the nation's power supply against being wiped out by two or three well-placed attacks. Power grids are important and need to be resilient to attacks of various kinds.
Solar farms in the far north will not generate as much electricity per square foot but will probably be worth it to reduce transmission losses and for the extended evening power to cool more southerly homes after the sun has set. Solar farms in the far south (from East to West) will maximize winter generation and hours of production.
Currently, fixed solar panels are orientated for maximum annual generation. As solar becomes a larger component of grid power, we will need more panels aimed westerly to deal with peak evening loads more efficiently. What this all leads to is more, smaller farms, not one massive "crown jewel" that would cripple the system if it suddenly went off-line for any reason.
Modelling the optimum strategy for placement of generation and storage capacity over time is a very complex problem with many solutions depending upon what the specific sub-goals are and will certainly involve installing and paying for panels that are only needed 10% of the time. That's how much cheaper generation is vs. long-term storage. The easiest and most productive way to solve for optimal solutions in a practical manner will be by using AI and Tesla is already getting their feet wet with their auto-bidder software. I suspect this effort will grow into ever larger modelling efforts used for the completely different purpose of grid planning.
The world is changing more rapidly than it may appear to the casual observer and there are heaps of money to made by those with a desire to understand these changes more fully and apply that knowledge appropriately as an investor. Tesla is probably the largest and most secure beneficiary of these changes but perhaps not the one with the highest returns. But remember, risk/reward is key in investing and not investing in non-productive companies is a big part of that. The future is difficult to see so probable growth is valued much more highly than more speculative growth.
Agree about distribution of solar.
There are about 8 billion m^2 of roofs suitable for solar in the USA and a similar size of parking lots (the size of parking lots is difficult to find). Together they could make up 80% or so of current electricity demand.
Wind tends to be strongest at night and in the winter (at least in Europe) and on a continent scale the wind is always blowing somewhere.
Continent scale grids, a combination of solar and wind together with batteries to smooth out peaks of supply and demand will enable robust electricity supplies for Europe and North America, probably China and India as well. Some hydro can also be used intermittently to smooth out differences between supply and demand.
Smaller regions, islands, etc. will need a larger amount of storage or go for nuclear.
Nuclear has three main disadvantages:
- firstly it costs more than a combination of solar, wind and storage and the cost difference is only likely to get larger over time,
- secondly common mode failures risk shutting down a large proportion at the same time (as in France),
- thirdly they are targets of war (as in Ukraine) and require a large trained staff at all times, the lifetime of a nuclear power plant is 70-100 years (including decommissioning). Many countries have been invaded, bombed, had civil wars or severe unrest over a similar time.
There are also demand side solutions, either reducing demand when it is high (e.g. temporarily turning off heating or cooling systems) or finding uses for excess supply (e.g. hydrogen, methane or amonia production).