It is usually stated that 80% of data has a spatial component. To be able to use this caracteristic at its full extent, companies and administrations set up geographical information systems. GIS were historically desktop tools, but evolved into distributed architectures, strongly web-oriented. In this context, PostgreSQL with its PostGIS plugin are key components for the data storage and manipulation layer. Be it for cartographic visualization, custom spatial data manipulation, geo-BI or geomarketing, PostgreSQL/PostGIS answers the needs on par or better than its main competitors, namely Oracle Spatial, SQL Server or MySQL. Based on ISO and OGC norms, PostGIS let the user deal with simple geographical features (points, lines, polygons), or more complex ones. Spatial data indices lead to high performances to query those geographical features. Numerous operators and functions allow plenty of geometric processing capabilities : creation, modification, feature relationship computations (intersections, proximity…), and even more complex processes, such as geometry aggregation, convex hull, geometric simplification... PostGIS also provides export functions, facilitating systems interoperability. JSON export makes it easy to write web applications based on OpenLayers, KML export gives a direct Google Earth visualization, GML aims at applications based on OGC standards (e.g. TinyOWS). PostGIS is currently becoming an OSGeo project, which will increase its quality and organisation. The next 2.0 branch will be full of new features, such as rasters, routing, geodesy, 3D support and more.