<p>Physics-based numerical models are essential for understanding atmosphere–space interactions. These include mechanisms that perturb the upper-atmosphere dynamics and energetics associated with terrestrial weather (impact from below) and solar-geomagnetic input (impact from above). The relevant physical phenomena range from global and synoptic scales down to the kilometer-scale of mesoscale dynamics. This manuscript discusses the impact of fine-scale processes in the ionosphere–thermosphere–mesosphere (ITM) region initiated from above and below, emphasizing the current status and perspectives of their realistic representation in whole-atmosphere and thermosphere–ionosphere models. The possible pathways to address challenges in realistic predictions of the ITM and its fast variability on global and regional scales are examined. They include three powerful modeling and data analysis techniques, namely nesting, nudging, and assimilation, that need to be carefully implemented for the highly dynamical ITM domain. Reasons for potential problems to constrain the high-resolution ITM predictions by data can be attributed to sparse and diverse observations performed in different layers and the increased complexity of the multi-scale ITM processes that need to be taken into account and properly coupled (e.g., dynamics, radiation, chemistry, and electrodynamics). The multi-scale ITM dynamics are dominated by tides, gravity waves, and small-scale eddies. Constraining these with assimilation and/or nudging algorithms requires rapid data analysis to properly adjust the global and regional diurnal and sub-diurnal oscillations to match space-borne and ground-based observations.</p>