Yesterday, I turned on my PC and saw a message on CAM (hardware monitoring app from NZXT) that my 1080 FTW has been "updated to new specs". I remember going to BIOS to downclock the CPU and overclock it again (I raised it to 4.5GHz on an i7 4770k) to experiment with certain games that were bottlenecked by the CPU. After I did that and started playing Dying Light, the game froze/black screen'd and the GPU fan was at 100%; I had to do a hard reset. There was also an instance (only once) where CAM displayed a "GPU fan has been disconnected" message, and Afterburner confirmed it when I saw 0% fan. 
 
Anyway, I researched the problem online and tried all solutions. I made sure that the power pins/connections are connected properly (GPU pins, 24 pin..,etc). My 750w PSU is new and a high-end model from Corsair (RM750i), and I double checked again using CPUID HW monitor and other monitoring devices. The problem is related to the overclock on my GPU; I thought it was the manual voltage curve but it happened again with a linear one. I reduced the overclock and only got drive crashes after 30-60minutes in The Witcher 3 instead of black screen+100% fan, but then it happened again and I had to downclock more. 
 
Could the overclock be the reason behind it? I don't remember that ever happening to me before, and it only happened after that CAM message and me messing with BIOS--I might have accidentally changed something. Right now, I am thinking of resetting the motherboard BIOS, or just try gaming without the overclock and see if it is truly the overclock that is causing it (I will settle down with a low overclock if that is the case). Could it be Afterburner glitching? Power issue? Defective GPU/going bad? I don't remember hearing about this issue due to an overclock, but could it be something new with the sensitive 16nm chips? 
 
I will keep on updating the threads until my system is stabilized. I appreciate any help and further explanation of the safety mechanic that causes the whole system to act like that (especially GPU fan). I don't want to RMA the card because I imported it; I don't know if I am eligible and I don't want to go back to the 780 and wait weeks for my card to be fixed (not sure how can I send it to EVGA via international courier service). 

ksgnow2010
Are you monitoring your system temperatures?  CPU and GPU temps?
 
Do you have adequate cooling to overclock the CPU?
 
Do you have adequate air flow inside the case?

Yes; I am monitoring them in real time via OSD; I double checked with the HW monitor software too. I have a HUUUGE Noctua NH-15 fan that keeps the CPU under 60c with heavy load, and the 1080 FTW is between 60-70c (highest is 75c with OC). The temperatures were good in the Witcher 3 even after playing for an hour straight (until driver crash because of my unstable OC). The motherboard temps are all good; I never had temperature issues before. As for my case, it is NZXT phantom full tower--it has good airflow and a lot of fans. The problem doesn't happen at idle, but only in demanding games when maintaining high clocks likes 2050-2088MHz (my card is average at overclocking). 

Remove all OC's and do some testing.  If your issue goes away, you'll have your answer...

Sounds like your cpu overclock caused the problem.