A couple things come to mind. One is that Twitter was originally(?) written in RoR, but they eventually switched to a different technology, because it didn't scale effectively. That was their story, anyway. I'm sure Rubyists would disagree that that was the real culprit.
Re. your point that enterprise software dev. teams don't do well with dynamic languages, I thought this commentary by Robert Martin of Object Mentor was interesting.
"What Killed Smalltalk Could Kill Ruby, Too"
http://www.youtube.com/watch?v=YX3iRjKj7C0Martin's main point, which I think he quoted from Ward Cunningham, was that what killed Smalltalk was "it was too easy to make a mess." Martin said that this could spell the doom of Ruby as well, though it hasn't so far, which is something good to ponder. Cunningham added that C++, for example, had the advantage that if you make a mess, it makes it very painful, very quickly. So programmers are motivated to be disciplined, and not to make a big mess in it. Whereas in Smalltalk, you could go on making a mess for a while, and not feel that affected by it. By the time it becomes onerous, and really starts to hurt, you've created a huge, incomprehensible mess! Another part Martin got into was that even though TDD was invented in the Smalltalk community, it didn't gain widespread adoption in it. So the messes became so prominent that they drove people away from it. He feared the same thing would happen with Ruby.
It's interesting hearing this angle, because another narrative I'd heard was that Smalltalk got killed because it didn't "get with" the internet, and Java sucked up all the talent that had been using Smalltalk before. Maybe it was a combination of both.
