The second-generation of supersonic civil transport has to match ambitious targets in terms of noise reduction and efficiency to become economically and environmentally viable. High-fidelity numerical optimization offers a powerful approach to address the complex trade-offs intrinsic to this novel configuration. Past and current research has proven the potential of supersonic aircraft shape optimization but lacks deeper insight on final layouts. This work partially fills this gap by investigating RANS-based aerodynamic optimization for both supersonic, transonic and subsonic conditions. We perform single and multi-point optimization to minimize the drag over an ideal supersonic aircraft flight envelope and assess the influence of physical and numerical parameters on optimization accuracy and robustness. Leading and trailing edge morphing capabilities are introduced to improve the efficiency at transonic and subsonic flight speed by relaxing the trade-offs on clean shape optimization. Benefits in terms of drag reduction are quantified and benchmarked with fixed-edges results. We observe how the optimized airfoils outperform baseline reference shapes from a minimum of 4% up to 86% for different design cases and flight conditions. The study is then extended to the optimization of a planar, low-aspect-ratio, and low-sweep wing, using the same schematic approach of 2D analysis. We investigate the influence of wing twist alone and twist and shape on cruise performance, obtaining a drag reduction of 6% and 25% respectively as the optimizer copes with both viscosity and compressibility effects over the wing. Preliminary results for 3D multi-point optimization suggest that the proposed strategy enables a fast and effective design of highly-efficient wings, with drag reduction ranging from a minimum of 24% up to 74% for cruise at different speeds and altitudes. The benefits of RANS-based aerodynamic shape optimization to capture non-intuitive design trade-offs and offer deeper physical insight are ultimately discussed and quantified.