Considering where the difficulty will go we're looking at 1 year to ROI.
Yes, if you were starting from scratch the ROI would probably be a year or more but everyone's situation is different. My current hardware paid for itself in July so if I sell my 20 S5s for something like $300 each I would have $6000 towards 5 S7s. That would make my initial investment about $3500 with 24,300 GH/s, 6050 Watts, and power at 0.10. If you plug those numbers in it shows 94 days to break even. Even if difficulty made a huge jump in the next 3 months it would look like maybe 188 days to break even for me. Like I said, everyone's situation is different.
Not really. You could just sell those machines and call it quits. Every purchase needs to stand on its own feet or there's no point to it.
On the flip side, I don't believe 16nm or 14nm processes will improve efficiency much beyond what the S7 can deliver. And 8-11nm is 3-4 years away. We might be looking at a miner that will be competitive for a couple of years after ROI is reached.
yeah I agree the 16nm and 14nm still has lots of kinks to work out with efficiency, i know intel is battling that now at some of their FAB's with the 3D tech.. uses more power than its worth