I'd be interested to see those numbers. Most of the "easy" hydro locations in the world are already taken so we can't just extrapolate the as-built costs of existing capacity.

Note: Hydro power is different than pumped storage. Hydro needs a river and a huge lake etc. Pumped storage can simply create a new lake on top of a mountain and pump water up there.

Colorado for example has a lot of high altitude lakes that are not used. Ideal locations need to be close to high grid producers or users otherwise you need to build a lot of transmission lines.

Further, current usage is designed for peak power output not total storage, if the goal is 10's of TWh we would want a lot of huge reservoirs with minimal pumping capacity each. Where current designs are focused on peak output which is the expensive part. So, massively increasing storage capacity would likely reduce costs per GWh stored.

Anyway, these numbers are going to have a lot of give based on goals. The point is the US government can make this problem go away for ~2 billion / year which means it's not a significant issue.

Sure, I understand that pumped storage is different technology but the reservoir + head infrastructure required for both is pretty similar and that's what costs money. The turbines would be designed differently since you want dispatch-ability rather than peak output and the intakes would be a lot lower because otherwise you can't actually use most of your capacity but otherwise you're still talking about massive civil engineering projects either way.

Not only that, but the best sites have probably already been taken and if you want to build a massive amount more (by a factor of 50 or so) then you'll get into more and more marginal sites.

Even in countries with a lot of lakes available near elevation change, blasting and tunnelling into rock to get the water from the high to the low reservoir is not cheap.

Assuming that you build one full day's worth of storage (10TWh), for the sake of convenience we can convert this to power and say that we need 416 GW. You say that the expensive part is focussing on peak output, I actually disagree (because I think the civil engineering costs dominate) but let's put a 50% discount on capital costs at the end to account for it.

Doing a little digging online, cost ranges for building pumped hydro go from $600 - $2500 / kW. Let's take the lower number and further halve it to account for what you said about energy maximising storage being cheaper than power maximising.

If we need 416 GW, that'll cost $208 billion US dollars to build.

To be honest I went into this thinking that the cost would come out to be ridiculous but I'm not actually sure that is an outrageous cost. It's only a fifth of the US annual federal discretionary budget.

If those cost assumptions actually added up which I'm not sure they do.

Edit: Hydropower generally builds a huge dam to support a large head of water piled on top of it's self. Pumped storage replaces that huge column of water with hill so you only need a relatively small amount of water to store a lot of energy.

Anyway, I still think your numbers are a little high as Solar and wind are never going to have zero output over a day which reduces peek demands significantly. Existing dams can also shift production to meet different sessional needs and act as backup power production. Still, let's call a ~200 billion investment enough to solve grid storage.

Further, current pumped storage can make money though arbitration. So, private investment subsidized by say 2 billion / per year or 80 billion over 20 years seems to make this a non-issue.

"Easy" hydro locations refers to hydroelectric power generation.

But the person you replied to was referring to pumped-hydro power storage, which is a totally different thing for a different purpose, and is much more flexible.