The aerospace industry is at a turning point: environmental concerns, legal frameworks and new energy sources mean that the industry needs to explore a different structural design space for composite aircraft configurations. Yet this is not readily possible using the slower experimentally-heavy design pyramid followed by industry in the past. The above scenario makes a compelling case for numerical structural design of very large integrated composite aircraft structures, but this problem is intractable. My vision is that structural design of very large composite structures can be enabled by a new simulation paradigm: I propose that, during the analysis, CAE models of very large structures adapt in real-time the scale of idealisation as required (adaptive multiscale), adapt in real-time the configuration of the structure (adaptive configuration), and where intelligent algorithms work at the back-end (HPC cluster) to extract high-value data from over 1 Tb output databases as they are being built. This paradigm will enable numerical design of very large integrated composite structures, thus having a significant impact on the emergence of much-needed new aircraft configurations.