One of my customers is using it. I also use it for a side project.

It is a robust graph database that seems to be optimized for exploring graphs from any given starting node. This is different from RDF graphs with SPARQL queries that are highly indexed and optimized for queries. There are also more hybrid systems like AllegroGraph that are both optimized for queries and graph traversal from a starting node.

There's also Stardog (stardog.com), from the makers of the Pellet OWL reasoner, which just entered this space. AllegoGraph is pretty cool, one of the best examples of commercial LISP success, and they have a pretty rich ecosystem around AllegroGraph (Jena, Sesame, etc support).

I have tried out two beta versions of StarDog, which is also very nice. If you ask, I think that they will provide you with the beta system.

already on the beta list ;)

Yes, we're using it for Whybase, and open sourced Bulbs, a Python persistence framework for it (http://bulbflow.com).

guess this is just the big names: http://neotechnology.com/customers/ but I lurk their forums and there are plenty apparently