The performance of a capacitive deionization (CDI) device is governed by complex relations between the electrode material properties, fluid velocity and fluid distribution within the device. In order to maximize fluid (water) interaction with the electrodes, the relationships between fluid flow and electrode material properties are explored here to develop novel CDI architectures which reduce the pressure drop, improve surface utilization factor and improve the electrode salt adsorption capacity. Using activated carbon cloth (ACC) as the electrode material, the pressure drop across the CDI device is quantified with respect to flow scheme (flow-between and flow-through CDI modes) used. Computational fluid dynamic (CFD) models are developed to study and optimize the fluid velocity and distribution in order to minimize the device fluid pressure losses. The model predictions are verified by constructing the conceptualized CDI devices and correlating the theoretical and experimentally obtained pressure drops, salt adsorption capacities and fluid flow parameters. The results indicate that up to 60% reduction in pressure drop and similar to 35% increase in specific salt adsorption capacity can be achieved by simple changes to the input-output port architecture of the CDI units. The results describe a method to considerably lower energy consumption in commercial CDI devices.