SQL# contains just about everything I'd want to do that isn't already in T-SQL.

It think SQLCLR has yet to have its day. I firmly believe that it has potential for developing a statistics toolkit and that requirement has yet to come to the fore.

We did some work with SQLCLR and service broker some time ago but it didn't catch on.

Personally I think that where there is separation between the various roles of a DBA and the roles of developers you won't get SQLCLR adoption. If you had developers with a strong database skillset then blending the two worlds together would be much more likely.