Worldwide glacier retreat outside the two large ice sheets is increasingly tangible and the associated ice-loss has dominated the cryospheric contribution to sea-level change for many decades. This retreat has also become symbolic for the effects of the generally warming climate. Despite the anticipated importance for future sea-level rise, continuing glacier retreat will affect seasonal freshwater availability and might add to regional water-stress in this century. Here, I envision a novel self-consistent, ice-dynamic forecasting framework for global glacier evolution that will lift the confidence in forward projections for this century to new heights. For the first time, each glacier on Earth will be treated as a three-dimension body within its surrounding topography without using any form of geometric simplification. The heart of the framework is the systematic utilisation of the rapidly growing body of information from satellite remote sensing. For this purpose, I intend to pass on to ensemble assimilation techniques that transiently consider measurements as they become available. This will streamline and increase the total information flow into glacier models. In terms of climatic forcing, global products will be replaced by regional forecasts with high-resolution climate models. Moreover, a more realistic representation of the local energy balance at the glacier surface is pursued that ensures multi-decadal stability in the melt formulation. The envisaged 3D finite-element modelling framework also allows a direct integration of iceberg calving, which is, on global scales, an often-unconsidered dynamic ice-loss term. To this day, a key limitation of glacier projections is the poorly constrained ice volume and its distribution at present. Here, I put forward a promising remedy that builds on multi-temporal satellite information to calibrate a state-of-the-art reconstruction approach for mapping basin-wide ice thickness on virtually any glacier.