Solar farms are ~0.8 MW / acre that adds up fast. Many current nuclear powerplants are similar energy density when you consider all the space inside the fencing.

Energy density is more a problem of materials. If you start talking GW hours it’s ~1GW from a ~38 ft tall and 48 meter diameter tank, even at $1,000 per ton it adds up. On top of that you need steam turbines etc.

From the description, it sounds like a giant Peltier device, not a steam turbine (otherwise why would they care about having a cold tank?). Solar farms also have a ton of space in between turbines so I’m not sure that metric is relevant. Additionally, since this device uses very safe materials (salt and coolant), there’s no reason you couldn’t dig a really big hole and shove the device inside...it would also provide some nice natural thermal insulation. When it comes to storing energy far from urban areas, cost is a much higher priority than energy density.

It uses gas compression through staged compressors to make the hot side, and gas expansion through staged expanders for the cold side, which makes a loop.

Think of it like a scaled up, reversible refrigerator.

I'd like to know if they also have a 'medium temp' tank, since if they do, they can keep the temperatures of the hot and cold tanks constant, and just change the amount of molten salt in them. That in turn means the compressors can be fixed compression ratios, which will dramatically increase efficiency.

To follow up here, having read the paper, they have two 'medium temperature' tanks (for the hot and cold side materials, which are different). The turbines do indeed work at constant input and output temperature, and at constant flow velocity and therefore wattage.

The paper suggests changing the pressure of the working fluid to adjust load, but I suspect it is hard to design a turbine to be efficient at a wide range of working pressures, so instead they'll end up with many parallel turbines, and just switch some off to get less energy produced.