Robotics would seem to be a big area of research. Using genetic programming to evolve good control algorithms, and even good functional form factors seems like the way things are going.

Further to that, I think there is a role for OR in artificial intelligence, in the sense of human-equivalent intelligence.

It seems online advertising, marketing, and recommendation engines are really popular as of late.

I also expect government bodies, policy makers taking OR analysts assistance in setting priorities and allocating funds.

Green energy solutions involve efficient combination of many disparate heterogeneous power sources like wind turbines in different locations, small hydro, micro hydro solar and conventional. Combining these efficiently is an OR task.

Also on the consumer end optimising the use of all this energy will be vital. This could be minimising traffic or smart timing of appliances or loads of other things.

So optimising energy production and consumption.

Retail is my pick if you are looking at commercial opportunities. Read Marshall Fisher's lecture on 'Rocket Science Retailing'.

As a grad student I certainly appreciate this question :)

I agree with adamo. Recommendation Systems, Collaborative Filtering Algorithms, Social Network Analysis are all running on original ideas from the field of Operations Research. The problem is, all of the new and exciting fields are interdisciplinary (CS, stats, OR, even civil and systems engineering)

I also like to argue that "Green Supply Chains" and green systems are going to be huge in the near future.

Medical procedures will be huge. I watched "Another Day: Cheating Death" on CNN, and learned new CPR techniques ignore mouth-to-mouth, and only use chest compressions. I don't remember the stats (OR sin), but the amount of lives it saved was amazing. It is a great show, and opened my eyes to the possibility of common practices in medical procedures being improved.

After reading "Catching the Network Science Bug" and the last chapter of "The Development of Social Network Analysis" I thought that social network analysis (especially now that data on networks with millions of people (or other agents) are available) would be an area for OR.

Definitely medical applications, both on the logistics of medicine delivery (nurse scheduling, etc.) and on the handling of medical procedures (e.g. optimal placement of radioactive pellets in prostate cancer).