The **Global Entity Reference System (GERS)** is a universal framework for persistent geospatial entity identification launched by [[Overture Maps]] in June 2025. GERS assigns stable 128-bit identifiers to 3.7 billion real-world features globally, solving the data integration problem in [[Spatial Data Science]]. Instead of complex [[Spatial Data Join|spatial joins]], datasets can link through simple table joins using GERS IDs as foreign keys. The system uses [[H3]] hierarchical [[Spatial Index|spatial indexing]] combined with entity identification, creating globally unique identifiers that maintain stability across data updates. GERS eliminates the "hidden tax" of geospatial conflation that often costs more than data licensing. Unlike proprietary systems (Google Place IDs) or community approaches (OpenStreetMap IDs that change with edits), GERS provides enterprise-grade stability with open licensing. Data is distributed as cloud-native [[GeoParquet]] files (~500GB per release) accessible through [[DuckDB]]. Common applications include multi-temporal analysis, real-time data fusion, and [[Machine Learning|ML]] feature engineering where stable entity identification survives model retraining cycles.