High demand for Text Mining tools? But why?

I got this email from TextMinr.com:

   At the moment, we are seeing extraordinary
   demand for our beta. To manage the growth of
   the user base and scale without melting down,
   there might be a wait of 3-6 weeks before we
   can send you an invite. We are very sorry for
   the wait, but you will get your invite as soon
   as possible, and we operate a "first come,
   first served"-policy, so your place in the
   queue is assured.

This is good news but I'm surprised. This means people see great value in text mining but why? It's not intuitive how much data can be scrubbed from text. How do you even know it's doable. OK, someone is making money out of Twitter sentiment but that's done. What will you make?

Another question is if TextMinr.com is up to the task. Text mining is not like any other data mining. It MUCH more difficult. Google prediction API is just silly because there is no control over feature engineering, feature selection, instance selection, or model selection. This leaves a lot of untapped market for predictive modeling startups IF (big if!) the demand is for profit not for fun or not for academics.