My takeaway from this article is a little different. Thermal loss is a function of three things. First, thermal difference: A hot water tank is at around 50C

My takeaway from this article is a little different. Thermal loss is a function of three things. First, thermal difference: A hot water tank is at around 50C / 120F, but it could be higher. Let's assume a 10C / 50F basement temperature. Second, R-value: Dividing the temperature difference to compare loss rates. For example, an R-value of 20 that insulates a 40 degree difference will be as effective as an R-value material that insulates a 20 degree difference, all other things being equal. R-values can go as high as R45, but typical wood+insulation walls are about R15. Pure wood is around R1 per inch thickness. Third, surface area of exposure. This is pretty self-explanatory. But if you put all these ideas together, then making a heat battery becomes basically a matter of raw scale. Say we design a cube storage tank. It is governed by these equations, where L is the side-length of the cube. R = InsulationR * InsulationLayers IA = InsulationArea = 6 x L^2 TIC = TotalInsulationCost = IA x InsulationUnitAreaCost * InsulationLayers VS = VolumeStorable = L^3 WVHC = WaterVolumetricHeatCapacity = 4.2 ( Joules / K ) / cm^3 MTD = MaxTemperatureDifference = MaxTankTemp - RoomTemp TES = TotalEnergyStorable = MTD * WVHC * VS EL = EnergyLost = Time * AverageTempDifference * IA / R Cost = ( TES - EL ) / TIC Basically I think this is a great idea for power storage if that hot water is need anywhere in the semi-near future. The larger the pools of water you have the cheaper it gets, economies of scale aside.