• Reporting on versioned data (that is, with configuration management enabled in RM and QM applications as well as the GCM application).  If you are following our release currently under development, you'll see that we've added a bunch of new data in TRS feeds (more on that below).  Both JRS Report Builder and RELM can use LQE with Configuration Management as a data source.
  
• Doing light-weight data integration across tools using the open specification Tracked Resource Sets (TRS), which are like news feeds of change events.  This provides a practical, incremental approach to data integration -- instead of big, expensive application integration projects.  Once the data from multiple tools are accessible via the LQE, one can write interesting queries that effortlessly cross tool boundaries, in a sense "unlocking" the data from their traditional tool-bound silos.  This is what RELM provides (including interesting multi-hop visualizations that help with understanding a complex system under development or analyzing change). The Eclipse Lyo project includes a TRS SDK. We've seen people create adapters for custom in-house tools and repositories, SAP applications, PDM/PLM applications, and other non-IBM data sources.
• As an internal data store for the CLM tools, for example the validity service behind the innovative validity/suspect capability in the release we are working on now uses an instance of LQE for its storage.

All that said, I don't see the data warehouse going away.  In addition to the Cognos Adapter, many dashboard widgets get their data from the data warehouse. Along with Tiago I expect over time teams will create a higher percentage of their reports from the JRS Report Builder, since it's so easy to use.