This is often repeated, because it is true. Just asserting it is false with no evidence is not very convincing.

Let’s compare solar growth with pumped storage growth. In 2010, US had 0.9 GW of installed solar power generation capacity. In 2020, it has 19 GW of installed capacity. That’s 2100% growth. In the same time period, pumped storage has grown from 21.5 GW to 22.8 GW. That’s 6% growth.

When should I expect to see significant growth of pumped hydro storage? I am willing to bet $500 that by 2030, pumped hydro generating capacity will not grow to more than 40 GW, while solar capacity will most definitely double by then. Will you take the other side of this bet?

It would be obviously foolish to build out storage before there is enough renewable generating capacity to charge it from. Until then, money is correctly spent on renewable generating capacity that actually displaces CO2 emission.

And that is, in fact, what is being done.

The only error is in spending less on renewable generating capacity than the looming catastrophe demands. I recommend you put your $500 there, instead.

So I take it you won’t take the other side of the bet, then? Just as I thought.

I have told you where you may send my winnings.

Hydro reservoirs seem to cost in the $50-90/kWh range. I don't see that going down as digging and pouring concrete are pretty cost optimized already.

Although a holes don't tend to wear out and can be used many more times than a battery, I'm not convinced low cost abundant-material battery options won't eat their lunch if we keep dragging our feet on the renewable buildout.

I guess if we were really committed to the idea we could use all that Uranium to make some very cheap holes and have it actually contribute to zero carbon energy rather than being a myth, but I'm not convinced the fallout is worth it.

Pumped hydro reservoirs do not, as a rule, require concrete, or overmuch digging, unlike your typical hydro generation dam, which people seem often confuse them with. (But as with most things, you can spend as much as you like, and some have.)

It is wholly possible that, in the fullness of time, some battery chemistry will undercut pumped hydro.

The main problem with chemical batteries is that the cost is linear with capacity, where pumped hydro cost tends to per square root of marginal capacity (for dike construction), but also costs per watt for the turbine-and-pump(s), with a fixed up-front cost for the penstock.

Have you got some examples of reservoirs in the 10s or 100s of GWh range in non-preexisting hydro projects that beat $50/kWh? If not then asserting that costs scale with the square root of capacity is a massive indictment rather than an argument for it.

The square-root scaling law means quadrupling the water storage only doubles the length of the perimeter dike. It is a good thing.

Dikes are as mature as any technology still in use: they predate writing. We don't need to guess what they cost. We have, similarly, well beyond a century of experience with penstocks and kinetic waterwheels. Pumps have been in use for some time, too.

> The square-root scaling law means quadrupling the water storage only doubles the length of the perimeter dike. It is a good thing.

Only if there is a project of reasonable size that has a manageable cost. Otherwise the corolloray that smaller dikes cost more implies that only a small handful of perfectly placed megaprojects will work. Responding with vague rhetoric for something you claim will work when there are hard numbers for finished projects using other technologies makes you look no better than the fission shills.

Snowy 2 blows fission out of the water, even being over budget by the usual fission ratio. But it had a watershed and it wasn't greenfield.

Are there any actually existing off river PHEL projects (or adequately sized reservoirs) with real budgets that actually got finished to compare?

California, in particular, has a century of experience with alpine reservoirs, penstocks, and kinetic water wheel generation; and have recently retrofitted pumping.

A century ago they built earthen dams up there using earth-moving equipment with parts operated by cables on pulleys, because hydraulics were not mature. The roads are still used today, mainly for recreation, and are execrable, but sufficed.

This is more vague hand waving. But it did at least lead to an (unfinished) project with a budget.

Eagle mountain is a brown field off river PHES project. Is it fair to say brown field projects are no more expensive than green?

It has 430m of head which is a fairly good site and a grade of about 30% which is excellent (so cost of power is minimized). Most cost grade A sites should be worse than this.

Cost: $2.5bn

Power: 1.3GW

Capacity: 18GWh or 14hr

Cost $138/kWh

Assuming half is the power infrastructure then cost of capacity is about $70/MWh. The levelised cost of a hole is about as close to zero as you want, but O&M costs $10 to a few tens of $ per MWh for existing hydro so there is no reason to think O&M would be cheaper than replacing a >5000-8000 cycle cell if it could match up front costs.

Compare to $200/kWh for current commercial iron/iron flow or historic lows of lithium (which should be indicative of an upper bound on sodium ion as the manufacturing is similar). CATL and Natron are claiming around $60-90/kWh is achievable for SIB by 2030, and Form energy (less believable but still probable) are claiming an eventual lower bound for Fe-air of $20-50.

Seems fair to say raw cost per kWh should favour chemical batteries in many areas by 2030 and LCOS should be on par in 2050s. Cost of power already favours batteries and colocation should favour them further by reducing transmission costs and curtailment.

Then there are electrolysers which have a high cost per use and per power but vastly lower (effectively free) cost of capacity.

Seems like fairly strong evidence that PHES is a poor fit in most areas unless started right now. Can you find a better example?