PS: The real advantage of hydro is you can invest in capacity or peak load as more or less separate things.
5.5 years is about the lifespan of servers in a datacenter. What's wrong with managing energy storage the way we manage datacenters? We don't expect servers to last 20-50 years. And no doubt, as battery tech improves, the lifespan will improve too. But the infrastructure to get good storage needs to start somewhere. Run them as datacenters for now while waiting for battery tech to get better.
Don't forget computer investment has long operated under the assumption of rapidly declining prices. With that slowing data centers may go to a 10+ year cycle.
I agree with you on hydro, gravitational storage is relatively simple to do. But it takes space. Batteries could make more sense for microgrids.
Electric car batteries can endure thousands of full cycles before their useful capacity drops down to 80%. That figure can be as high as tens of thousands, depending on who you ask.
Suprisingly, for LiMn, temperature and age have a much greater effect on lifetime than cycles.