Assuming utilities will do the stupid, expensive thing leads to failed analyses.
Sounds like an Elon Musk solution. Make it bigger, taller, make more of it, and somehow it’ll bend basic physics facts to become the ground breaking disruptor technology of the next century.
Mineshaft storage is very, very efficient, but has an upper limit on energy per mineshaft determined by its depth and how much weight fits in it. The cost of the winch apparatus is amortized over that amount. More mineshafts means more total storage.
Pumped hydro has very, very low cost per Wh stored, but kinda high per W in and out. But thousands of pumped hydro systems were already built, almost, so seem free; they just needed pumps added.
Synthetic H2 and NH3 are very, very cheap per Wh stored. W out uses existing turbines, so seems cheap. W in depends on your electrolysers, which are improving fast.
In batteries, cost per Wh stored has been large, but is different in different chemistries. In a better chemistry it is lower. Flow batteries have lower cost per Wh stored, for any liquid chemistry, but higher per W in or out; and need plumbing and pumps.
Details matter. Numbers matter. Dumb choices make dumb designs. Dumb designs can be improved on.
Drawing conclusions based on dumb choices is just dumb.
Musk is dumb, but just barely smart enough to hire smart people and, sometimes, let them be smart. So, Hyperloop was just dumb, no matter who works on it. Starship is kinda smart, depending on what you can think of to do with it. Mars colonization is dumb. Starlink could possibly be smart, depending on a lot of details.
A wholly new capability, like radically cheaper mass to orbit, can often enable smart ideas. Super-expensive mass to orbit was good enough for weather satellites, planetary probes, and geosynch COMSATs.
That one had the winch and motor / generator at the sea floor, and a big rack of floats. It was, of course, too expensive. This is a common feature of energy storage ideas: people seem to love to promote the most stupidly expensive version of a good idea, and then give up when it is shot down for that.
Certainly, any scheme where you put complicated electromechanical stuff on the sea floor, it will be expensive. And likewise, where the expensive part is only usable with one "unit" of storage. Or anyway it just costs too much or needs perfect weather or something. (E.g., Energy Vault (NRGV) is stupid and their investors will lose their shirts.)
But there is usually a simple variation that lacks these problems.
That is why vertical motion in a disused mineshaft is attractive. A single 10,000t weight hanging from a winch has just one moving part, the winch, and cable all in tension, with no preparation of the shaft sides; everything is concentrated at the top. The mineshaft provides isolation from environmental disruption.
You get much the same effect with a float attached to a cable down to a pulley augered into the sea floor; thence to an onshore winch. The cable and pulley are exposed to sea (or lake) water, but you get the advantage that the winch may be shared among as many cable reels as you like, with only simple clutches. A 20m diameter float displaces 4000 tons of water, and the sea floor may be much more than a km deep.
(In principle you could share a winch drive among lots of mineshafts, but they are rarely close enough together. You could bore more, but that drives up cost. The world is absolutely perforated with disused mineshafts.)