I think that there are several factors to take into consideration for the VRM temp:
1) case ventilation and overall temp in it - in my case I have a 20cm fan which blows air on my video cards - see Corsair Carbide 500r case
2) overclocking - if you put power at let's say 110% you will clearly force the VRMs (voltage regulators) - here at the moment I have "stock" voltage
3) fan curve - here I keep it on aggressive, which should help
I am not sure if one can measure VRM temp, in the normal operation, not by removing backplate and putting some thermal sensors on it an so on....
And, if I know correctly, the "exploded" scenarios happened only with GTX 1080 FTW and not with GTX 1070 FTW... even if both cards suffer from the same "wrong" thermal design.