This dataset presents the AI-quifer global raster database, a harmonized collection of 38 georeferenced maps describing the geological, hydrological, hydrogeological, geomorphological, and climatic setting of the world's land areas and continental shelves between 60° S and 90° N. All layers are provided as GeoTIFFs in EPSG:4326, with a common spatial extent (−180 to 180° longitude, −60 to 90° latitude) and pixel sizes ranging from ~0.0083° to 0.1°. The database includes global geology (terrane type, sedimentary terranes, rock age, erosivity), seafloor properties (sediment thickness, porosity, crustal age, bathymetry), and geophysical fields (magnetic anomaly, gravity model, crustal type, distance to plate boundaries). It further integrates surface and coastal information such as land cover, proximity to the shoreline onshore and offshore, palaeo-coastline age and lowland exposure since the Last Glacial Maximum, as well as catchment-scale descriptors (mean elevation, slope, area, erosivity, and a catchment-maturity index based on area–Strahler order relationships).The database also compiles a suite of hydrological and hydrogeological variables, including global land water storage anomalies from the GLWS 2.0 product (mm equivalent water thickness), groundwater head and water table depth, groundwater recharge, aquifer thickness, specific yield, hydraulic conductivity, and changes in recharge between present day and 30 ka BP. River-network–related layers comprise global flow accumulation (linear and log₁₀-scaled), slope-limited flow accumulation and slope maps derived from Blue Earth bathymetry/topography, and coastline-centred Strahler-order heatmaps and catchment statistics based on HydroSHEDS river networks. All layers are consistently processed (reprojection, resampling, masking, truncation and log-transforms where appropriate), and fully documented through a structured metadata table that records data sources, DOIs, licenses, processing level (raw/modified/derived), units, no-data values, and data type (categorical, ordinal, continuous).The AI-quifer database is designed as a foundational input for global-scale analyses of coastal and offshore groundwater systems, especially for machine-learning and process-based studies that require co-registered global predictor variables (e.g. offshore groundwater occurrence, coastal aquifer vulnerability, sedimentary basin characterization, and coupled land–sea water-storage dynamics). By combining widely used open datasets (e.g. USGS global geology, GEBCO/Blue Earth bathymetry, GLWS 2.0, PCR-GLOBWB2, HydroSHEDS, GFZ products) into a single, consistently processed framework, this dataset facilitates reproducible, data-driven research on marine and coastal hydrogeology and can be readily reused in other global Earth system and environmental applications.