Accurately quantifying groundwater recharge remains a persistent challenge despite the use of advanced hydrological models, physically based methods, water budget approaches, and isotopic tracers. These traditional techniques are limited by uncertainties in input parameters, scarce field data, and methodological non-uniqueness where different combinations of parameters can produce similar results, making it challenging to accurately determine actual recharge conditions. Such limitations hinder their reliability as standalone tools for recharge estimation and our ability in water planning and management. Recent advances in satellite remote sensing, particularly through the Gravity Recovery And Climate ExperimentGravity Recovery and Climate Experiment (GRACE) (GRACE), offer new possibilities for addressing these challenges. As discussed in this chapter, GRACE provides large-scale observations of terrestrial water storage changes, enabling complementary insights into groundwater processes, including groundwater recharge even at sub-regional or local scales when downscaling techniques are applied. This chapter explores recent progress in applying GRACE data to groundwater recharge assessment, emphasizing the integration of satellite gravity observations with numerical modelling and groundwater budget approaches. Special attention is given to emerging machine learningMachine learning and computational techniques that enhance the spatial resolution of GRACE-based recharge estimates to provide a holistic perspective on water balance. As the coarse spatial resolution of GRACE data places significant limitations on its applicability for estimating recharge in localized or significantly smaller groundwater systems, new capabilities to generate high spatial resolution GRACE data using machine learning are crucial to help understand groundwater processes, including recharge on a significantly smaller catchment scale that underpins groundwater management. The chapter demonstrates the feasibility and value of combining downscaled (or high spatial resolution) GRACE data with traditional methods or in situ data to better characterize recharge dynamics and close existing knowledge and data gaps in the assessment of groundwater processes.