Root system architecture is classically seen as the geometrical side of plant strategies to capture soil water and nutrients. In our quest to optimise water extraction under drought, target architectural phenotypes have been proposed and several paths (phenes) to achieve them have been identified. Nevertheless, recent progress in phenotyping, biology and modelling, indicate that the contribution of root system architecture in the capture of water could be more complex than what we have considered. A first avenue ahead of us relates to the temporal dynamics of root system architecture. Progress in phenotyping and modelling helps us capturing rules constraining the variability of root elongation. A second avenue relates to the plasticity of root development. Current research in model and crop species has identified novel mechanisms of root response to soil water heterogeneity, including hydrotropism, hydropatterning and xerobranching. This knowledge should help us to build more realistic and less empirical models of field root architecture. A third avenue relates to the multi-scale nature of the soil-root hydraulic continuum. The most recent models of soil-root hydraulic architecture now include rhizosphere properties, aquaporin expression and cell-wall decorations at the cell level, root anatomy and root development, which are within the reach of phenomics. The presentation reviews our progress in these avenues and questions whether these represent new opportunities to achieve efficient capture.