There is not a single report of VRMs overheating with the FTW series.
There are reports of faulty VRMs which EVGA has confirmed and given an estimate about 3%-4% of the cards being affected BEFORE 30/8.
What to do?
On the backside of your EVGA box, there is the serial number. Use this serial number to check via the GuestRMA
http://www.evga.com/support/guestregister.asp when your card has been produced and delivered to the reseller.
If before 30/8 there is a 4% chance you have a card with a faulty VRM.
If after, there is no chance.
So, you can decide if you want to take the risk or not. A faulty VRM does not mean it WILL combust. There is the much bigger chance it will simply not work properly and give you black screens and freezes with the GPU fans spinning up to 100%
This is NOT related to the overheat spin that 2nd class pc media outlets put on the skewed findings of ONE review, which is already called false by EVGA, not condoned by Nvidia because of wrong testing conditions and critizised by individuals with at least basic knowledge in temperature measurements to be inaccurate. Again: There is not a single report publicly available that can pinpoint any issue of the FTW to overheating.
Now, the thermal pads EVGA offers for free for cards not having them and are installing on new batches of these cards are to lower the VRM temperatures even further, to put your mind at ease and to be within specs comparable to other brands with VRM cooling via thermal pads. The VRM operate within spec accordingly as EVGA claims. Given that the company puts out roughly several ten thousands of cards each fiscal year and we see maybe over a dozen issues reported here in the forum about a card being affected by the faulty VRM or some issue with the cooling (pads not touching the chips, I got you guys covered :P) is within acceptable tolerances given that brands like Asus had to pull the plug on an entire line, the Strix o80 and had issues with micron VRAM on other models.