In this paper, a novel geometry is proposed for evaporators that are used in Supercritical Organic Rankine Cycles. The proposed geometry consists of successive plenums at several length-scale levels, creating a multi-scale heat exchanger. The channels at the lowest length-scale levels were considered to have their length determined by the thermal entrance-length. Numerical simulations based on turbulent flow correlations for supercritical R134a and water were used to obtain performance indicators for new heat exchangers and baseline heat exchangers. The relationship between the size of the channels at successive levels is based on generalization of “Murray’s law.” Using the data on pumping power and weight of all metal structures, including that of all the plenums, the total present cost was evaluated using a cost model for shell-and-tube heat exchangers. In addition to the total present cost, the data on overall thermal resistance is also used in identifying optimal heat exchanger configurations. The main design variables include: tube arrangement, number of channels fed from plenum, and number of rows in the tube bank seen by the outside fluid. In order to assess the potential improvement of the new evaporator designs, baseline evaporator designs were created from the new evaporator designs by simply removing all the internal plenums, resembling a traditional shell-and-tube configuration. Consistent with geothermal applications, the performance of new heat exchanger designs was compared to that of baseline heat exchanger designs at the same flow rates. The present costs for the new multi-scale HXs were estimated to be approximately ¼ of their corresponding baseline HXs. This four fold reduction in present costs of the new HXs was found to be attributed mainly to high operational costs, as evidenced by the much higher pumping power required by the baseline HXs, while the heat loads and the total mass of the alloys for the new HXs were very close to those for the baseline HX. The significant cost savings in the new HX designs compared to those of the baseline HXs, at similar heat load performance, are a strong evidence that the new HX architectures proposed in this paper are valid alternatives to traditional HX designs.