Is it compressed air? The article says, and I quote: > To store energy, excess electricity is used to pump water out of the sphere, creating a relative vacuum

Is it compressed air? The article says, and I quote: > To store energy, excess electricity is used to pump water out of the sphere, creating a relative vacuum. To release energy, we open the valve: the water, pushed by the external pressure, rushes into the sphere and turns the turbine, producing electricity. If you take air at ~50 bar and release it into the atmosphere and try to extract energy from the water replacing it spinning a turbine, you are throwing away most of the energy stored in the compressed air. In any case, the math works out for pumped hydro. Ask Google: (4.5m)^3 * 4/3 * pi * 500m * 1 g/mL * 9.8m/s^2 in kWh This gives 520 kWh, which is consistent with the article’s claims. edit: One can ballpark the storage capacity of compressed air storage, too. Assuming isothermal compression (which would be a nice ideal case and is not easy to achieve unless one compresses very slowly), the work is nRT • ln (volume ratio). nRT = PV measured at any point in the process, which is conveniently exactly the calculation above: 520kWh. For the pumped hydro model, I lazily assumed that they pumped all the way to vacuum, which is obviously wrong (some water would boil), but it makes almost no difference. But here we need to compare the actual pressures, and the pressure ratio (equivalently volume ratio) is around 50 between sea level and 500m deep. So multiply by ln 50 to get around 2MWh. But that’s the actual work done in a perfect isothermal process. In the real world, the starting and ending states will be around the same temperature, but the process will be far from isothermal, so a good deal more than 2MWh will be used to compress the gas and a good deal less will come back out.