Yesterday (argh, the day before yesterday) I went to a SAI conference on nosql. In Belgium, SAI is a non-profit organization for IT people which focuses on knowledge sharing.
The conference that day was on nosql. The presentation given by OuterThought was very good and offered a nice introduction to the "new types of database architectures" that are being actively developed as we speak.
Although the use of these nosql databases within KBC (where I work for) is limited (I'm not aware of any application that is already using this technology) it would be plain wrong to discard the technology as "too immature". With the recent developments that we face in the IT industry, applications are nowadays quick in adopting such new technologies and I suspect that off-the-shelf applications will soon come with such nosql database technology as part of the solution.
For large enterprises, this does face some (hard?) challenges: how do you control your network usage (some of the technologies are easy to use, but hard to tune), how do you design your architecture, where is your data, how can you ensure that you do not "lock in" into a single nosql technology (i.e. how do you ensure interoperability and migrations between technologies), do you still need SAN-based replication or will you now let the technology handle this for you, etc.
So yes, the nosql technology is nice to look into (and definitely something to follow up on) but make sure you don't introduce it in your organization without thinking about the entire integration and management aspect.