The observed degree of thermodynamic imperfection of existing power plants is explained based on a steady-state power plant model the irreversibility of which is due to three sources: the hot-end heat exchanger, the cold-end heat exchanger and the heat leaking through the plant to the ambient. While maximizing the instantaneous power output, it is shown that in addition to Curzon and Ahlborn's optimum temperature ratio there exists also an optimum balance between the sizes of the hot- and cold-end heat exchangers. A graphic construction for pinpointing the optimum location of the power plant on the absolute temperature scale is presented. The efficiency (first or second law) is maximum when the total investment is split optimally between the external conductance and internal thermal resistance built into the power plant. The efficiency data on existing power plants fall in the domain anticipated theoretically. Some of the trade-offs revealed by the theory are illustrated further by the analysis of an ideal Brayton cycle power plant. © 1988.