We present the first 1992‐2020 record of Greenland Ice Sheet (GrIS) mass balance derived from multi‐satellite Ku‐band altimetry. We employ an empirical approach as an alternative detailed to radar‐propagation modeling, and instead convert elevation changes observed by radar altimetry into mass changes using spatiotemporal calibration fields. This calibration field is derived from a machine learning approach that optimizes the prediction of a previously published mass balance field as a function of ice sheet variables. Our mass balance record shows a GrIS contribution of 12.1±2.3 mm sea‐level equivalent since 1992, with more than 80% of this contribution occurring after 2003. Our record also suggests that the 2017 hydrological year is the first year in the 21st century which, within uncertainties, the GrIS was in balance. Overall, the 28‐year radar‐derived mass balance record we present highlights the potential of the method to provide operational mass balance estimates derived from multi‐satellite Ku‐band altimetry.