Faith + Reason = Truth

This is not a full review, just a quick thought based on the specs of Excel 2007.  Spreadsheets and databases are different technologies with different uses, each with abilities to do things the other can't.  Every corporate environment uses Excel, as well as Microsoft Access.  Often times, however, the main (and frequently ONLY) use of Access is for dealing with datasets too large for Excel to handle with its 65,000 row limit.  So analysts get the data into Access, run some quick filters to get it to a manageable size, then take the filtered dataset back to Excel to analyze.
This is no longer necessary with the Office 2007 version of Excel, because Excel 2007 now has a 1 million row limit, plenty large for most dataset analysis.  Now, you'll need LOTS more RAM to run Office 2007 efficiently and to work with million record spreadsheets, but this should knock a broad swath of current, limited Access users right off the map.  The users with true, multi-table requirements to link data sources together will still need to stick with Access, of course (or FileMakerPro, whose latest release includes VERY cool web integration for mapping, among other goodies), and eventually may have some real alternatives with web 2.0-based apps such as DabbleDB, which are currently limited by their inability to work with datasets larger than several thousand records (but what these web apps CAN do in terms of ease-of-use and creative uses are well worth an exploratory effort if the datasets aren't too large).