It'll be interesting to see how much these will cost. 4 Teraflops of single precision calculation is very impressive, however if you need double precision, the speed is not quite as impressive at 400 Gflops - although still very good compared to my laptop with 20 Gflops :)

The main drawback with these is that you have to use CUDA, which certainly takes a while before you wrap your mind across. I played CUDA for a while but considered it is too much effort for something which is very specialised and might not become mainstream. Still there seems to be many people using CUDA and with lots of research roles - see the CUDA forums.