sorry, i wasn't talking about design feasibility, rather the market feasibility... what any designer would need to know is the worth of his effort... if i develop a system and people still prefer GPU's over fpga rigs, then why would i want to waste my time?... in short if i could such a system in the market, how many people would want to buy it?
Double check your numbers for 5W and 900 Mhashes/sec... If you still can produce at these specs, you'd have lots of interest...