I gave up my FPGA project - even if it's cheaper than ASICminer - because at current difficulty (and difficulty increase rate) the ROI rised to 6-8 or even more months. I have better investments than that :/ I'll leave these two PCBs I already made running and forget the whole thing.
6-8 months is a great ROI, actually...
Well, it's very risky 6-8 months. Current lead-time for Artix-7 from the manufacturer is 2 months. The difficulty in 2 months? Currently it's rising at 1-1.5M per day, so it should be around 70-100M then. So the ROI will be counted in years. And who knows what happens during that time? The ASIC manufacturers will keep their arms race, the difficulty will rise even more and the ROI will never come. Not thinking about actual income.
P.