A grid is concerned with the integration, virtualization, and management of services and resources in a distributed, heterogeneous environment that supports virtual organizations across traditional administrative and organizational domains. Spatial data infrastructures (SDI) aim to make spatial data from multiple sources available and usable to as wide an audience as possible. The first SDIs of the 1990s followed a top–down approach with the focus on data production and centralization. In recent years, SDIs have seen a huge increase in the number of participants, necessitating a more dynamic bottom-up approach. While much research has been done on web services and SDIs, research on the use of data grids for SDIs is limited. In this paper an emergency response scenario is presented to illustrate how the data grid approach can be used as a decentralized platform for address data in a dynamic SDI. Next, Compartimos (Spanish for ‘we share’) is presented, a reference model for an address data grid in an SDI based on the Open Grid Services Architecture (OGSA). Compartimos identifies the essential components and their capabilities required for a decentralized address data grid in a dynamic SDI. It deviates from the current centralized approach, allows data resources to come and go and node hosts to grow and shrink as necessary. An address data grid in an SDI is both a novel application for data grids as well as a novel technology in SDI environments and thus advances the mutual understanding between data grids and SDIs. In conclusion, additional research required for address data grids in SDIs is discussed.