I have found two very informative videos with possible explanation why all these video cards are failing.
 
Steve from Gamer's Nexus reached out to EVGA about it. EVGA claims that out of 1 million cards sold only 200 failed (lol sure ok!).
 
In another video it's discussed that basically EVGA got a batch of defective "components" from a manufacturer. These components went into these lemon video cards which fried.
This video discusses the major changes and steps EVGA took once they identified there were faulty components into making their 1080ti FTW3.
 
My question is why we these dangerous cards which can possibly fry and burn out with some starting small brief fires have not been recalled? They are obviously health and safety hazard to us. Other manufactures are forced by government to recall their product and rectify the issue. When Samsung Note 7 was frying they were ALL recalled. Samsung took responsibility for their lemon Note 7 despite having found that the fault was with out of spec batteries coming from one of their suppliers. Why isn't EVGA responsible like other companies and do the right thing? Instead of resolving this issue with replacing these fire bombs with redesigned FTW3 cards EVGA will happily send you another fire hazard ticking bomb (if you are with in warranty). They claim only 200 cards out of 1 million are faulty this should be very simple solution to continue keeping their customer safe and happy.

Even though these videos focus mainly around 1000 series problems I believe that the bad components started with end portion of 900 series production.
 
EVGA's VRM Thermals Not the Killer of Cards - Final Test
https://www.youtube.com/watch?v=mpxQaSjQclo
 
EVGA Goes Insane with 1080 Ti FTW3 PCB (VRM Analysis)
https://www.youtube.com/watch?v=4UURdOEMrCA

the_Scarlet_one
I think you misunderstand a very pertinent point, these articles were about a very specific card.

The 200 to 1m scale was to show a very small percentage of cards received bad those components. .02% is the amount.. It wasn’t EVGA saying “exactly 200 cards have won the lottery to be a guaranteed failure”.

 
I very much understand where EVGA comes up with these numbers. These number are downplayed and don't reflect reality in my personal opinion. These is a result of small percentage of people who attempted RMA on their cards and weren't DECLINED due to the 100 rules and regulations EVGA has on the RMA page. Most people probably just threw the card in the bin and bought a new one. 
 

the_Scarlet_one
Samsung was a large company and was having massive issues with batteries exploding.

 
So EVGA is excused from doing the right thing and removing these fire bomb video cards because they aren't as big as Samsung?
 
Samsung was just ONE example of many examples "if you google you will find many more" where companies were forced to recall products for safety reasons. Big companies and smaller companies!

the_Scarlet_oneYour card was a 900series card, which wasn’t part of the articles you posted.

 
Thank you for pointing that out because I didn't include that in my post stating that I think the issue BLEW UP with the 1000 series all over the internet and finally hardware sites started making posts/videos about it. My card was bought January 1st 2016. That's 3 months? before 1080s were on the market. That's why I think the older faster 980s which were manufactured at the end of the line probably have the same faulty components are the early 1000 series cards. I stated this therefore I don't understand why you need to point this out again?

the_Scarlet_oneI suggest using google to look up “GPU caught fire”
Guess what, that is an MSI card. Capacitor and vrms have had failures on all gpu’s. The only reason EVGA gained any notice was because reddit decided to make it become noticeable.

 
Seriously? You found a random MSI card burning out and u think this is relevant in this discussion? Are you that desperate to show us that other brands fail not just EVGA?
Look closer at that video. Look at that budget CX430 PSU with those garbage cables. Look at the motherboard heat sinks. Look at the stock Intel CPU cooler. You got some ghetto rig which looks like my computer from 12 years ago with some inferior components burning out the card. The video is posted in 2016 but was it shot in 2016? No description explaining the model of the video card. 
If I had my video card running in a budget build computer like this guy I would expect to have hardware frying. 
I got people coming to me asking me to build them a PC for under $1000. I warn them that they might have issues with a computer like that. When I cut costs and buy budget PSU I tell'em it might burn out in 1-2 years. I expect problems. When I have a $5000 PC like me and picked out some the best parts for it I expect it to be problem free. Only reason anything might fail is when companies like EVGA try to cut their manufacturing costs by going with some shady lower grade components.
 
Go google videos about buying hardware parts in china markets. You have same part coming in 2-4 different variants  of quality. They ask you how good do you want the quality of the product and off course it well cost more.
 
The reason EVGA gained notice is because of the large volume of their cards burning out. You want to impress me go find me equal volume of ASUS 900/1000 Strix cards frying like that. 

Wolfpack Hybrid

the_Scarlet_one
I think you misunderstand a very pertinent point, these articles were about a very specific card.

The 200 to 1m scale was to show a very small percentage of cards received bad those components. .02% is the amount.. It wasn’t EVGA saying “exactly 200 cards have won the lottery to be a guaranteed failure”.

 
I very much understand where EVGA comes up with these numbers. These number are downplayed and don't reflect reality in my personal opinion. These is a result of small percentage of people who attempted RMA on their cards and weren't DECLINED due to the 100 rules and regulations EVGA has on the RMA page. Most people probably just threw the card in the bin and bought a new one. 
 

the_Scarlet_one
Samsung was a large company and was having massive issues with batteries exploding.

 
So EVGA is excused from doing the right thing and removing these fire bomb video cards because they aren't as big as Samsung?
 
Samsung was just ONE example of many examples "if you google you will find many more" where companies were forced to recall products for safety reasons. Big companies and smaller companies!

the_Scarlet_oneYour card was a 900series card, which wasn’t part of the articles you posted.

 
Thank you for pointing that out because I didn't include that in my post stating that I think the issue BLEW UP with the 1000 series all over the internet and finally hardware sites started making posts/videos about it. My card was bought January 1st 2016. That's 3 months? before 1080s were on the market. That's why I think the older faster 980s which were manufactured at the end of the line probably have the same faulty components are the early 1000 series cards. I stated this therefore I don't understand why you need to point this out again?

the_Scarlet_oneI suggest using google to look up “GPU caught fire”
Guess what, that is an MSI card. Capacitor and vrms have had failures on all gpu’s. The only reason EVGA gained any notice was because reddit decided to make it become noticeable.

 
Seriously? You found a random MSI card burning out and u think this is relevant in this discussion? Are you that desperate to show us that other brands fail not just EVGA?
Look closer at that video. Look at that budget CX430 PSU with those garbage cables. Look at the motherboard heat sinks. Look at the stock Intel CPU cooler. You got some ghetto rig which looks like my computer from 12 years ago with some inferior components burning out the card. The video is posted in 2016 but was it shot in 2016? No description explaining the model of the video card. 
If I had my video card running in a budget build computer like this guy I would expect to have hardware frying. 
I got people coming to me asking me to build them a PC for under $1000. I warn them that they might have issues with a computer like that. When I cut costs and buy budget PSU I tell'em it might burn out in 1-2 years. I expect problems. When I have a $5000 PC like me and picked out some the best parts for it I expect it to be problem free. Only reason anything might fail is when companies like EVGA try to cut their manufacturing costs by going with some shady lower grade components.
 
Go google videos about buying hardware parts in china markets. You have same part coming in 2-4 different variants  of quality. They ask you how good do you want the quality of the product and off course it well cost more.
 
The reason EVGA gained notice is because of the large volume of their cards burning out. You want to impress me go find me equal volume of ASUS 900/1000 Strix cards frying like that. 

I agree with you, don't even waste your time.  that's why i went with asus for MAINBOARDS and GPU'S.
power supplies are about the only thing i buy anymore from evga had problems with both hardware from
evga.not one asus product has ever burnt out on me, so next buy should be a asus card i would hope.
 
 
edit Cool GTX - removed the double quote (same post quoted twice)

My 980 ti made it 3 years and 3 months. It popped 2 weeks ago and I am trying to wait on a 2080 ti to replace it with.

I have to say that I've had the exact opposite results from eVGA products.
 
Every eVGA card I've ever bought I still have, and they still work (with the exception of the original 8800GTS, which I upgraded via Step-Up to a GTX 260 (that eVGA was nice enough to provide me with an SSC version of).
 
My 970, as the primary card since I bought it, is still going strong - just shy of 3.5 years later.
 

Part Number: 04G-P4-3979-KB
Part Desc: EVGA GeForce GTX 970 SSC ACX 2.0

---

Place Purchased: Best Buy
Purchase Date: 04/26/2015
Date Registered: 4/26/2015 5:29:35 PM
 
You have to realize that electronics are fragile.  You have 0 proof that the problem was not exacerbated by the line voltage in your household, or a bad ground that allowed voltage from nearby lightning strikes to come in on the electricity lines and cause damage to your components, or any one of well over a THOUSAND different things that can (and *DO*) happen on a daily basis, even in well industrialized countries like Canada (where I lived as a child for a few years) as well as my native US, let alone whatever other, lesser-industrialized nation my username inspires in whatever scenarios.
 
I've had similar discussions with other system builders I know.  One that tickles me is a friend of mine in the NE part of the US - swears up and down backwards that he will never, ever buy Seagate drives, that they are crap.  I've never lost a Seagate drive due to inexplicable failure - the only 2 (out of something approaching 20 drives in 25+ years of system building) I've ever lost were a pair of Seagate barracuda 500GB 7200.11s - and they were taken out (as was the motherboard and my PSU) by a lightning strike quite literally right outside my house.  In fact, the monitors I had connected to that system eventually also failed on me - but I never forgot that tiny little (but significant!) fact - they were connected to a system that, for all intents and purposes, took a (filtered) but nearly direct lightning strike.
 
Furthermore, today's electronics are not handcrafted like an Italian sports car, or a Rolls Royce - they are mass produced.  And when they are mass produced, you need mass shipments of components to produce in volume.  And if a batch of those components are bad, then guess what?  Yep, you've guessed it - a batch of the products that that batch of components are in tend to go bad as well.
 
Your experience (and others, as well) show that not everything is perfect in the world of eVGA electronics.  But, just as you'll never buy anything from them again, I plan to continue to buy from them - I've had very little problems from them, (aside a niggling issue where my old 560 Ti 448 Core was always detected by PrecisionX / PrecisionX OC software as a 760).  The only limiting factor might be my disillusionment with Intel CPUs, and if eVGA doesn't shift to support AMD processors, then that will make the decision on the motherboard for me automatically.  So, I plan to continue to buy from them, as they have never done me wrong.
 
Here's hoping that you will find an OEM that you like well enough that you can suffer a failure or two.  because, I can guarantee it - something else you buy that is electronic will, at some point, fail.
 
Welcome to the world of mass production.

Sorry to heAR about these issues

Hello,
 
I had cards fail before. 780 GTX, 980 Ti and 2x 1080 Ti FTW3. In all the previous cases I had to send the cards back ... from Qatar !!! Because I bought them from the US.
 
In the 780 GTX case, the RMA was good. It was cross ship by the way and I got a full refund when they got the defective card, not sure if it was refurbished or new. For the 980 Ti , it was within 3 years of warranty, at this time , it seems there was no 980 ti stock so the replacement was 1070 , and guess was , I was running SLI and the changed my second working 980 Ti with a 1070 SC to match the original replacement card. For the 1080 ti FTW3 , in SLI too, a card failed, I got a bad replacement , and the next replacement was a brand new card. And I paid a lot of extra shipping costs because I bought the cards from the US and had them shipped to Qatar. But this is not Evga's fault. I buy hardware internationally and willing to take the risk of such scenarios.
 
All my RMA experiences was fast AF. I didn't even had to upload any screenshots or whatsoever. Just shared the troubleshooting that I did to identify defective cards and immediately RMA was issued. On all my RMA experiences.
 
3 things I learned from my previous experiences:
 
1. Parts fail. (even if they are new and expensive)
2. Evga customer support is excellent.
3. You will not find same support anywhere else.