Accurate specification of high-latitude electrodynamic forcing remains one of the major challenges in storm-time ionosphere-thermosphere modeling. Current physics-based forecasting models typically rely on empirical descriptions of auroral precipitation and electric fields derived from historical observations. While these empirical models perform reasonably well under quiet and moderately disturbed conditions, they often struggle to capture the highly dynamic high-latitude forcings during intense geomagnetic storms. Observation-constrained field-aligned currents (FACs) offer an alternative approach for specifying high-latitude electric fields. Previous studies have shown that FAC-driven simulations can successfully reproduce large-scale high-latitude electric fields and many key features of the ionosphere–thermosphere response during moderate-to-intense storms. However, their ability to accurately capture the complex electrodynamics and multiscale ionosphere-thermosphere coupling associated with intense and extreme storms remains an open question. In this talk, I will present recent efforts to model storm-time ionospheric and thermospheric responses using observation-constrained FACs. Through several case studies, I will discuss what aspects of the ionosphere–thermosphere system can be successfully reproduced, where important discrepancies remain, and what these successes and limitations imply for the future space weather forecasting.