ACTUAL Failure rates for RTX 2080 from my investigation

Eb9922

n00b
Joined
Nov 23, 2018
Messages
3
Since word broke about issues arising with the RTX cards, I have been monitoring the situation very carefully, perhaps obsessively. A trend I have noticed reading through many threads, is a massive amount of conjecture, accusations of Nvidia/retailers lying about RMA figures, and a lot of numbers being thrown around as to what the actual RMA/defect rate is on many of these cards. Being as though I am the soon to be owner of a RTX 2080 (ROG STRIX OC), I have closely followed word on the 2080 specifically, and have done my own investigation as to what the failure rate could possibly be on this particular model. Admittedly, I do not have access to totally complete data, as no one other than Nvidia knows the total amount of cards sold, actual RMA rates, etc. But what I am attempting to do is use all the data that I do have available to me, and rather than just provide conjecture, instead provide my actual findings.

Findings on the RTX 2080
----------------------------------------------------
Newegg
135 reviews
1 fail EVGA reference board (non specific failure)
1 fail EVGA reference board (non specific failure)
1 fail EVGA reference board (possibly user error)
1 fail ASUS reference board (non specific failure)
Amazon
114 Reviews
1 fail MSI Custom Board (non specific failure)
1 fail MSI Custom Board (non specific failure)
1 fail ASUS reference board (non specific failure)

Failure rate= 2.81 %
---------------------------------------------------
Poll from right here on Hardforum
"Do YOU have Problems with YOUR RTX card?"
104 Votes
RTX 2080 received 1 Vote 0.7% Percent
---------------------------------------------------
REDDIT Results of the 2080 reliability Survey
119 RTX 2080 Cards submitted
115/119 cards working or 97.5%
Not-Working 3/119 or 2.5%

---------------------------------------------
Der8auer on RTX 2080/Ti RMA rates (AiB cards only)
Results as of 11/5/18
RMA RATES
2080 ~ 0.17%
2080Ti ~ 1.4%
It must be noted that according to Der8auer, he indicated that MORE 2080's have been sold than 2080ti's at this point, a figure he described as "approaching 4 figures".
So, more 2080's sold, and hugely less RMA's than the 2080ti.


-------------------------------------------------
FINDINGS ON RTX 2070
For the RTX 2070, Newegg currently has 58 reviews. Amazon has 45 reviews.
Of those 103 reviews, I found 21, yes, TWENTY ONE, reviews saying they have received a faulty card. That is a defect rate of 20.39 %.
I find this figure to be striking, and definitely raises some red flags to me.
------------------------------------------------
OTHER FINDINGS
1.Newegg currently has 55 reviews for the 2080ti. Amazon currently has 56 reviews, for a total of 111 reviews. Although this doesn't tell us the actual sales figures are, we can probably exclude the argument that more 2080ti's are in the wild versus the 2080. Evidence suggests that is probably not the case.
2. Of the failures that I did find of the RTX 2080, only 2 if I recall correctly, mentioned the word "artifacting". Some of the reported "failures" in those reviews I am skeptical of, as some were complaints of games crashing, no video signal coming from the GPU, etc.

Other things to note.

1. Founders Edition 2080's are not included in my data above, due to data like user reviews not being available. Therefore, I cannot ascertain as to how that would change any of my findings.
2. I am not claiming that the RTX 2080 is immune from the issues that the 2080ti is facing. I'm certain that once this thread is live, there are going to be users chiming in who have had issues with their 2080. I have seen some evidence that the 2080, in two reports, can display the same exact "Space Invaders" artifacting that is common with the 2080ti. I'm just simply providing evidence that this "issue", whatever it may be, doesn't seem to be plaguing the 2080 in the way that it is affecting the 2080ti, or what seems like, the 2070 to some extent.
3. Again, I am not presenting any of this data as "Absolute", I am just making the most likely conclusion I can with the available data.

As to what the true nature of this problem is with the RTX cards, truth be told, I don't have a clue. Initially, I thought that this might just be a case of the technology being pushed too far in the design of these cards, but seeing the same failures are happening on the 2070, it seems that's probably not the case.

Undoubtedly, there are people out there with cards that have issues, this is undisputable, and evidence that I have seen says that the RMA figure that Nvidia has given on the 2080Ti is questionable at best. But with that said, there is also a part of me that wouldn't be surprised to find that a large amount of these reports, complaints, etc, are in fact false claims made by individuals looking to fan the flames of this controversy. There is little doubt in my mind that, had these cards been released at a lower price point and had better performance than they do, these issues would not have been blown up to the extent that they have been.

That said, what have I learned by all of this? Truth be told, now I have more questions than answers at this point.
 
Last edited:
Those are interesting observations. To me, the failure rates seem to be in the expected range, I've been buying cards since the Riva TNT and helped launch the Geforce 3 even. I still lean heavily on nvidia hardware and have been purchasing and still purchase both their consumer level and enterprise equipment today. I think the 2080ti represents pushing the pricing envelope on the consumer level worse than any other GPU launch which is why it has generated a lot of attention.
 
Nice findings. I could believe it's 3% or in that range.

That is always why I question people that aren't buying. Most likely we are talking about 90% or more of cards with no problems.

And worst case, you wait a week and do an RMA. I wouldn't let that hold me back, that's for sure.
 
People go online to complain more than they do to complement.

Plus most of the time Amazon reviews are from the same kind of people that freak out over someone putting onions and or tomatoes on their burger, how dare they!

However, if you're going to increase the price on your products, you better have wiped your ass real clean as not to leave a shit stain, problem is like every other generation their cards are not perfect, but this time people are looking closer to fault nVidia (which they should).
 
People go online to complain more than they do to complement.

Plus most of the time Amazon reviews are from the same kind of people that freak out over someone putting onions and or tomatoes on their burger, how dare they!

However, if you're going to increase the price on your products, you better have wiped your ass real clean as not to leave a shit stain, problem is like every other generation their cards are not perfect, but this time people are looking closer to fault nVidia (which they should).

Dude, I almost spit out my soda reading this. You're totally right.
 
I don't see any reliable data here to lead to any conclusions. The survey can yield some preliminary generalizations however.

End of day, buy a card with a warranty you can easily use if needed with least hassle.

And yet the overwhelming mass public has drew the conclusion that massive amounts of RTX cards are dying left and right, solely based upon some owners that made some complaints in discussion forums.

I, on the other hard, have actually provided real data from user reviews, RMA rates from actual retailers, and polling's from those same forums that people are complaining from.

If the figures I provided in my post do not quality as "reliable data", may you enlighten me as to what would be considered "reliable data" in order to draw a conclusion? Does the official RMA rate provided by Nvidia qualify as "reliable data" to you?
 
And yet the overwhelming mass public has drew the conclusion that massive amounts of RTX cards are dying left and right, solely based upon some owners that made some complaints in discussion forums.

I, on the other hard, have actually provided real data from user reviews, RMA rates from actual retailers, and polling's from those same forums that people are complaining from.

If the figures I provided in my post do not quality as "reliable data", may you enlighten me as to what would be considered "reliable data" in order to draw a conclusion? Does the official RMA rate provided by Nvidia qualify as "reliable data" to you?

Your analysis is completely flawed. You are trying to show there are no problem with the RTX cards based on the 2080 card. Why didn't you do your analysis with the 2080Ti cards?

Retailers will not give you honest RMA rates of a product they are trying to sell you, neither will Nvidia.
 
Interesting read, I am curious to know exactly what the failure rate is just because it seems like it's very high at the moment for the 2080ti cards.
 
You can't trust the reviews of cards because not everyone writes a review (good or bad). You can't even tell what percentage of sales are reviewed (e.g. maybe 3% of total sales have a review attached to it). Generally, people with problems will tend to review things. At best its anecdotal and doesn't give you any useful data.

Best course of action is buy a card with a good warranty. (Or don't buy an RTX card at all due to Nvidia's pricing (y) ).
 
You can't trust the reviews of cards because not everyone writes a review (good or bad). You can't even tell what percentage of sales are reviewed (e.g. maybe 3% of total sales have a review attached to it). Generally, people with problems will tend to review things. At best its anecdotal and doesn't give you any useful data.

Best course of action is buy a card with a good warranty. (Or don't buy an RTX card at all due to Nvidia's pricing (y) ).

Agreed. No point in worrying about failure rates. Either step aside and wait until the bugs are worked out or just buy a good card with a good warranty (like evga).
 
And yet the overwhelming mass public has drew the conclusion that massive amounts of RTX cards are dying left and right, solely based upon some owners that made some complaints in discussion forums.

I, on the other hard, have actually provided real data from user reviews, RMA rates from actual retailers, and polling's from those same forums that people are complaining from.

If the figures I provided in my post do not quality as "reliable data", may you enlighten me as to what would be considered "reliable data" in order to draw a conclusion? Does the official RMA rate provided by Nvidia qualify as "reliable data" to you?

I have had my reservations on replying to this thread because even if you are unlucky as you suggest it is not fun ending up with a card you have to send back regardless of brand X or Y.
Having said that the failure rate at [H] somewhat makes me worry that it is something worse then expected. Which also makes a point about your thread and about the nonsense Gamers Nexus was spreading.
As reaper12 was saying that you will never get a honest answer and can you from Nvidia?
And why is that, why do you have to have people give anything but honesty they paid for your product (some say to much but Nvidia only sells top end hardware according to many many people) and you expect more a lot more.

When a company as Nvidia can't even be honest about GPP why should it be honest about failure rates but low and behold they will replace bad cards (isn't that great news everyone rejoice).

Your real data is also something that will be upsetting even more because you ridicule people as complaining while they have every right to be royally pissed off by this epic fuckup.
 
And another concern is with some people reporting having multiple cards fail on them. With low failure rates, this is something that should not be happening, but I've read multiple people running into this issue (even Mr. Bennett was unfortunate here). While I am one who jumped on buying a 2080Ti and haven't had any issues so far, I definitely know what it feels like to have a piece of hardware fail right out the gate. Never a fun time going through the RMA process at all. I can't imagine having to do it twice in a row.
 
The think that's most interesting to me is the people showing up with low post counts saying there's nothing wrong, so "Buy Nvidia!!!"

That's interesting. We'll see in a few months either way.

Either the problem is a nothing burger, or the shill accounts will have outed themselves, and we'll probably know both of these things just after Christmas.

There will be either tons of people still pissed off, while Shills are doing the "Iraqi Information Minister" impression, or the channel will have shaken out all the "Test Escapes" cards.

The odds of Kyle's two out of three cards going bad in short order is very troubling to me; if I was responsible for this debacle, I'd be hoping to get a new job somewhere before they fire me.
 
You come to a forum with facts and not FUD OP?
Oh you are either going to get ignored (because the data doesn't suit peoples agenda) or flamed for not jumping the bandwagon of "+25% failure rate"....facts are not important these days :/
 
You come to a forum with facts and not FUD OP?
Oh you are either going to get ignored (because the data doesn't suit peoples agenda) or flamed for not jumping the bandwagon of "+25% failure rate"....facts are not important these days :/

Rubbish, how can you make a claim about the failure rates of the 2080Ti based on the returns rates of the 2080? Because that's exactly what he has done.
 
The one thing I dislike about review, and please call me out if I am wrong, is that usually reviews for other items that might not be similar model are mixed in the list. Amazon for example, if you scroll through some of the reviews for one particular item show reference to something completely different. The numbers should be taken with a grain of salt, IMO. As to the 2080TI failure rate. I was one of those FE owners, but to Nvidia's credit they remedied the problem 3 days from letting them know.
 
Back
Top