They only make sense when you have large amounts of low carbon power that you want to store for long periods of time.

Usually, just turning down some fossil fuel plant is a better option.

Using the electricity via demand response comes after that.

A long wire to somewhere else that can use it and send you back some of their excess is next.

If you really can't use that power for anything lithium batteries are cost efficient as long as you regularly cycle them (e.g. daily solar peaks).

So you need to get into storing power for long periods of time before they have a niche. Which is on its way, but isn't here yet. So building more renewables makes more sense. (Do enough of that and eventually you'll want some more storage options).

Yes, demand for all kinds of storage will grow exponentially as renewable generating capacity approaches saturation: meaning you generate, in aggregate, as many joules in an average day as will be consumed in that day.

That happens over and over for each house that gets solar installed and the grid is not trusted, or is badly overpriced (as in many places with nukes).

The article has a link in it to: https://www.sciencedirect.com/topics/engineering/flow-batter...

If you scroll down there's a table 32.7 about Projects in China. A number of which, according to that table, are completed.