Solar Is Cheapest Electricity in History, U.S. DOE Aims to Cut Costs 60% by 2030

I think the future will be robust national/international grids, with a mixture of storage options (batteries/pumped hydro) to smooth out the intermittent nature of wind and solar.

Cynics always talk about the amount of energy storage required for solar as if you need to store 24 hours of energy for solar/wind to be viable.

I'd like to see numbers on having 1 hour of storage for peak demand, a robust national grid, and appropriately provisioned and placed solar and wind, taking the duck curve into consideration.

Even achieving just one hour of storage globally amounts to 2.5 TWh of storage. By comparison the entire world produces ~300 GWh worth of lithium ion battery annually. That leaves geographically limited options like pumped hydroelectricity, and solutions not yet deployed at any significant scale like hydrogen fuel cells, synthetic methane, thermal batteries, flywheels, etc.

Realistically we should saturate daytime energy demand with solar, and if there aren't any scalable storage options by then switch gears and proceed with hydroelectric where it's viable and nuclear where it's not.

Why would we use lithium ion batteries?

I would imagine the approach to store the energy would be to use the energy from solar panels to do work that can be used to produce electricity later.

For example, you could use solar energy to pump water back uphill to flow down through a hydro electric dam later.

Even if it isn't the most efficient, in the long run it would likely provide the best scalability and least long term environmental impact. Once you have the facility in place, the same water could be pumped uphill to flow back down a million times over with the only overhead replacing water lost through evaporation and maintaining the facility.

Am I missing something that makes such an approach unfeasible?