I realize that I am answering a paleo-question from a year ago, but I think it is a good question....
When it comes to career choice, it is vital to find the "right question" before trying to find the "right answer".
What I mean is you may NOT want to choose a career path because it is (or may become) "high demand". Sure, you'll want to choose something that makes you employable and that yields a sufficient salary and on the surface it sort-of makes sense that high-demand jobs will do that.
The problem is that with high demand eventually comes throngs of competition which is composed of wannabe's to super-elites and everything in-between. Unless you're talking about fields with very very long educational requirements, you're going to quickly find yourself in an ocean of similarly qualified candidates all looking for "the best" opportunities. Some will find their dream jobs but many will find themselves in unsatisfying career ruts after making a bad choice simply because it seemed like a good idea at the time.
Moreover, it is truly a guessing game to try to select what the hottest technology or career path is going to be in the future, or what will fall out of favor or be sustaining. It might seem that Java web development, for instance, is the most flush with opportunity for now and the near future, but what about the next things coming down the pipe. I would have sworn in 2003 that .NET was going to bury Java-- but guess what, it hasn't. Who is say what is coming next? Nobody really knows and it is dicey to make career plans based strictly on such reasoning (not saying either .NET or Java would be mistake).
Basically I am trying to say two things:
1. There is NO WAY to really predict what is going to be "next"
2. You may not want to choose a career path based on demand.
It is vastly more useful to find something that you're very good at, that you enjoy doing, and that puts you in a unique niche. Instead of trying to put yourself in the center of the bell curve and hoping for the best, find something in what people have called "the long tail"-- something that very few employers need but which is very valuable to those few employers.