One of the interesting things about hydro is that it's usually constructed to satisfy baseload power. In reality, shifting that over to a peaking plant requires relatively modest changes to system, a small fraction of the cost of an entirely new dam. You don't actually need the "Pumped" part of pumped hydro, you can just throttle normal hydro on and off if you have enough turbines (though for ecological & geomorphological reasons some minor downstream damming also helps). There wasn't any reason to install the extra turbines in the age of fossil fuels. They only take ~30 seconds to spin up, versus days or weeks for thermal plants.
> You don't actually need the "Pumped" part of pumped hydro, you can just throttle normal hydro on and off if you have enough turbines (though for ecological & geomorphological reasons some minor downstream damming also helps).
Yeah that generally can work for a week or so, not for entire season.
Imagine if we build hydro to full capacity and monetize the excess with data/compute centers and/or crypto mining. We'd have so much spare capacity for any residential or industrial purposes.
Datacenters and crypto are evidently a bad source for large-scale load variability because their significant capex demands that they be kept running 24/7.
This is often a problem with variable load schemes that do any useful work, rather than just dumping heat into something. It would be interesting to see a list sorted by just how thick they are with capital considerations. Aluminum smelting, hydrogen electrolysis, lots of other options on that list.
> their significant capex demands that they be kept running 24/7.
US companies are sitting on a pile of not yet connected GPUs because they have no place to put them. They would take 10 out of 24h if they could get 10. It's better than zero that they currently have for those GPUs.
Same with crypto. No point of running a rig if it's losing you money becaus electricity is too expensive at this moment. Regardless of capex sometimes it's better not to run it.
