Law Firm Investigates Class Action Suit Over Intel's Unstable 13th/14th Gen CPUs

[Dr_b_]

Dead/Dying CPUs, class action lawsuits, mass layoffs, suspended dividend.... AMD is totally in the rear view mirror right about now, amiright fellas?

 

[Crackong]

Losing this law suit and paying the fines might be CHEAPER than actually recall all the affected CPUs

 

[Minus Infinity]

I guess sacking 17000 people should give them enough for their legal fund.

 

[user556]

Arguably, every model affected has been permanently damaged to some degree ... and the lawsuit may even make such an argument. A full recall and replace program would make sense.

 

[Bwaze]

Arguably, every model affected has been permanently damaged to some degree ... and the lawsuit may even make such an argument. A full recall and replace program would make sense.

So all CPUs are damaged to some degree? A mid-August fix will magically eliminate all further damage, and prior damage will not shorten CPU lifetime or affect stability or performance to any significant level. Was your CPU damaged significantly? Then we might talk RMA.

 

[mkppo]

So all CPUs are damaged to some degree? A mid-August fix will magically eliminate all further damage, and prior damage will not shorten CPU lifetime or affect stability or performance to any significant level.

Damn you know a lot about a future microcode update, it's like words are coming out of intel themselves. Or you have a pretty expensive crystal ball..

 

[Bwaze]

I've seen enough large companies dealing with such issues over the years. The text I wrote above won't ever come out of Intel, because they would never admit that the whole range of CPUs are damaged to some extent - even small extent.

The text that will accompany microcode, bios releases will talk about certain motherboard setting damaging a small portion of CPUs - they know most of the processors out there aren't utilised hard at all, so they can safely bet there's tons of CPU in perfect health, and treat the ones with instabilities as rarities.

How rare? That's the neat thing, we don't get to find out.

Just look how confident are they of the longevity of their CPUs:

Intel Extends Warranty of Boxed 14th & 13th Gen CPUs By Two Years In Light of Instability Issues

"We stand behind our products, and in the coming days we will be sharing more details on two-year extended warranty support for our boxed Intel Core 13th and 14th Gen desktop processors."

So if you bought pre-built there is no extended warranty?

 

[JustBenching]

The text I wrote above won't ever come out of Intel, because they would never admit that the whole range of CPUs are damaged to some extent - even small extent.

Wouldn't that apply to other CPUs that were force fed excessive soc voltage and even though they survived, some damaged has been done already? At least with these intel chips you get a 5 year warranty, with those other cpus you don't.

 

[Chomiq]

Since it wasn't posted here before:
https://www.reddit.com/r/hardware/comments/1ei1zvm

 

[JustBenching]

Since it wasn't posted here before:
https://www.reddit.com/r/hardware/comments/1ei1zvm

And as I've said - nothing to do with rejecting rmas just because. Just read the thing you posted, intel claims that the cpus are probably not genuine and asked him to return them to microcenter, and that if he wants to proceed with the RMA he can - but if the CPUs are in fact found to not be genuine they won't RMA it.

 

[mkppo]

I've seen enough large companies dealing with such issues over the years. The text I wrote above won't ever come out of Intel, because they would never admit that the whole range of CPUs are damaged to some extent - even small extent.

The text that will accompany microcode, bios releases will talk about certain motherboard setting damaging a small portion of CPUs - they know most of the processors out there aren't utilised hard at all, so they can safely bet there's tons of CPU in perfect health, and treat the ones with instabilities as rarities.

How rare? That's the neat thing, we don't get to find out.

Just look how confident are they of the longevity of their CPUs:

Intel Extends Warranty of Boxed 14th & 13th Gen CPUs By Two Years In Light of Instability Issues

"We stand behind our products, and in the coming days we will be sharing more details on two-year extended warranty support for our boxed Intel Core 13th and 14th Gen desktop processors."

So if you bought pre-built there is no extended warranty?

Yep, i've seen a ton of microcode update over the decades. But how can you state that the microcode update will fix the issues and also not significantly cause CPU's damaged prior to the microcode any instability? There's just no way to tell that and entirely depends on how far the chips have degraded.

In fact Intel have admitted that all 65W+ are prone to higher than usual levels of degradation and that CPU's damaged prior to the microcode will have no fix other than RMA. So they have said it..

 

[Bwaze]

Yep, i've seen a ton of microcode update over the decades. But how can you state that the microcode update will fix the issues and also not significantly cause CPU's damaged prior to the microcode any instability? There's just no way to tell that and entirely depends on how far the chips have degraded.

In fact Intel have admitted that all 65W+ are prone to higher than usual levels of degradation and that CPU's damaged prior to the microcode will have no fix other than RMA. So they have said it..

It's very hard to prove that there is an issue over whole product release at all with certainty. Sure, Intel has lots if data, OEM partners have lots if data, but tech journalists, users? We have anecdotal data. People are reporting issues in forums? Every product has issues! When is too much?

Intel might target just that uncertainty. After the microcode fix there will be a long period when all issues will be questioned - was the damage done before microcode fix, and therefore irrelevant? Is it just normal CPU abuse due to overclocking? And after a couple of months it will be irrelevant, all the marketing will focus on upcoming product, best ever in all areas, even in reliability!

 

[evernessince]

Since it wasn't posted here before:
https://www.reddit.com/r/hardware/comments/1ei1zvm

Those aren't the only reports of that happening. Plenty of people on the Intel reddit have been reporting the same, at least the posts that haven't been deleted.

I'm going to be honest, I have zero faith that reddit actively goes out of it's way to ensure moderation isn't receiving gifts and compensation for biased moderation. There's no financial incentive for them to do so.

To top it off, Puget published numbers that show Zen 3 and 4 with higher failure rates than 13th and 14th gen but confusingly with 11th gen Intel having the highest failure rate overall.

Puget's numbers for even it's lower failure rate CPU generations are about twice the average and in general they make no sense. To me it screams low sample size, bad data, or bad SOPs on Puget's end as CPU failure rate absolutely does not vary anywhere near that much outside of this recent Intel issue. This data is definitely only relevant to pugent given no other source reflects their data.

 

[mkppo]

Those aren't the only reports of that happening. Plenty of people on the Intel reddit have been reporting the same, at least the posts that haven't been deleted.

I'm going to be honest, I have zero faith that reddit actively goes out of it's way to ensure moderation isn't receiving gifts and compensation for biased moderation. There's no financial incentive for them to do so.

To top it off, Puget published numbers that show Zen 3 and 4 with higher failure rates than 13th and 14th gen but confusingly with 11th gen Intel having the highest failure rate overall.

View attachment 357439

Puget's numbers for even it's lower failure rate CPU generations are about twice the average and in general they make no sense. To me it screams low sample size, bad data, or bad SOPs on Puget's end as CPU failure rate absolutely does not vary anywhere near that much outside of this recent Intel issue. This data is definitely only relevant to pugent given no other source reflects their data.

I've written something in the other thread just to explain this better. But in short, this is to be expected in chips not being run 24/7 and you have to factor in the fact that these 13/14th gen include a lot of CPU's that are not top of the line i9's which are showing the most degradation.

Also note the field failure rates, 13th gen is double that of 14th gen which shows these chips are failing over time. Ryzen 7000 that was launched before 13th gen has something like 0.25% field failure rate, so basically once they're deployed nothing happens to them. Which is pretty much what the server vendors are saying. Whereas those people running intel CPU's say they don't last past the 6 month mark and sometimes even lower.

Also I agree, no other source has the same data and the data itself has too many caveats. A better source would be a system integrator that has actually deployed server farms using these chips 24/7. They haven't published graphs, articles and such but many of them are saying the same thing - the higher end intel 13/14th gen CPU's have failure rates over 25%. All of them can't be wrong..

It's very hard to prove that there is an issue over whole product release at all with certainty. Sure, Intel has lots if data, OEM partners have lots if data, but tech journalists, users? We have anecdotal data. People are reporting issues in forums? Every product has issues! When is too much?

Intel might target just that uncertainty. After the microcode fix there will be a long period when all issues will be questioned - was the damage done before microcode fix, and therefore irrelevant? Is it just normal CPU abuse due to overclocking? And after a couple of months it will be irrelevant, all the marketing will focus on upcoming product, best ever in all areas, even in reliability!

That's a different topic though. My point was, there's no way to tell the microcode update will 'magically' solve everything. If you look at GN's vent at intel, it's apparent that they've been trying to fix this issue for a while as they've known about it. Till date they've been unsuccessful, but lets see what the new one brings. But again, your initial quote was neither stability nor performance will really be affected. My point is there's no way to tell other than actually see what they bring out because, judging by their recent track record, there's no guarantee..

 

[JustBenching]

I've written something in the other thread just to explain this better. But in short, this is to be expected in chips not being run 24/7 and you have to factor in the fact that these 13/14th gen include a lot of CPU's that are not top of the line i9's which are showing the most degradation.

No, those numbers are based on 700k and 900k.

Those aren't the only reports of that happening. Plenty of people on the Intel reddit have been reporting the same, at least the posts that haven't been deleted.

Intel handles 10s if not hundreds of rmas a day.Until now we have 3 reports, all of them very very serious. Let's check them out

1) First case, guy wanted to RMA, Intel said sure, then he changed his mind and asked for a refund. Intel told him to apply for a refund with the retailer he bought his CPU (obviously, that's exactly how it happens in the EU) but nope, he got mad and opened a thread to complain that intel wasn't refunding him, lol.

2) Second guy was told that if they don't have the same CPU available they will instead give him something better - he didn't understand what that meant somehow and he made a video complaining for 0 reason. Imagine intel telling you we don't have a 13500 - we will replace it with a 13600k and youd be crying about it,lol.

3) Third guy send a picture of his cpu die full of paste and whatever bot intel was using couldnt properly read the serial number. Of course he opened a thread complaining for 0 reason. Microcenter told him to clean the damn paste and resend the picture - he did - and intel went through with the RMA.

Those are the very serious reports we have of intel not honoring rma. We need to stay vigilant and on our toes - intel is denying us warranty

 

[mkppo]

No, those numbers are based on 700k and 900k.


Intel handles 10s if not hundreds of rmas a day.Until now we have 3 reports, all of them very very serious. Let's check them out

1) First case, guy wanted to RMA, Intel said sure, then he changed his mind and asked for a refund. Intel told him to apply for a refund with the retailer he bought his CPU (obviously, that's exactly how it happens in the EU) but nope, he got mad and opened a thread to complain that intel wasn't refunding him, lol.

2) Second guy was told that if they don't have the same CPU available they will instead give him something better - he didn't understand what that meant somehow and he made a video complaining for 0 reason. Imagine intel telling you we don't have a 13500 - we will replace it with a 13600k and youd be crying about it,lol.

3) Third guy send a picture of his cpu die full of paste and whatever bot intel was using couldnt properly read the serial number. Of course he opened a thread complaining for 0 reason. Microcenter told him to clean the damn paste and resend the picture - he did - and intel went through with the RMA.

Those are the very serious reports we have of intel not honoring rma. We need to stay vigilant and on our toes - intel is denying us warranty

Yes they are based on 700k as well, and not all are based on 900k which are the ones in question with accelerated degradation. Then take some time to read the other points before spamming this chart everywhere. There are multiple caveats.

Also, you seem to be hell bent on defending Intel when it comes to RMA. Well answer this question then, why were server farms denied multiple RMA's after their chip degraded? What are consumers who bought prebuilts going to do when their 1 year warranty expires? Without extending the RMA by 2 years which is pointless, maybe include tray CPU's to be RMA'd as well, wouldn't you agree? I mean, Intel literally had it on their page that tray CPU's are included, only for them to silently edit it later and remove tray CPU's. What is your opinion on that?

 

[JustBenching]

Yes they are based on 700k as well, and not all are based on 900k which are the ones in question with accelerated degradation. Then take some time to read the other points before spamming this chart everywhere. There are multiple caveats.

Also, you seem to be hell bent on defending Intel when it comes to RMA. Well answer this question then, why were server farms denied multiple RMA's after their chip degraded? What are consumers who bought prebuilts going to do when their 1 year warranty expires? Without extending the RMA by 2 years which is pointless, maybe include tray CPU's to be RMA'd as well, wouldn't you agree? I mean, Intel literally had it on their page that tray CPU's are included, only for them to silently edit it later and remove tray CPU's. What is your opinion on that?

I'm hell bent on providing facts. Those are the 3 popular cases with threads / videos on youtube and reddit about Intel "denying RMA". They are spammed everywhere, even though none of them have intel actually denying RMA.

But sure, if you prefer me to go along with the intel bashing sure, let's do it

What good is a 5 year warranty when Intel denies 100% of the RMAs, multiple users are complaining on reddit left right and center. Terrible company, ill only buy amd from now on.

Also, you seem to be hell bent on defending Intel when it comes to RMA. Well answer this question then, why were server farms denied multiple RMA's after their chip degraded? What are consumers who bought prebuilts going to do when their 1 year warranty expires? Without extending the RMA by 2 years which is pointless, maybe include tray CPU's to be RMA'd as well, wouldn't you agree? I mean, Intel literally had it on their page that tray CPU's are included, only for them to silently edit it later and remove tray CPU's. What is your opinion on that?

Who told you that prebuilds don't have or have 1 year warranty? Maybe it's a US problem? Cause in EU it's illegal, you can't even sell a product with only 1 year of warranty. 2 years is the minimum.

 

[chrcoluk]

There is all sorts of stuff been spread around the internet for a multitude of reasons, I watched this video earlier today, and it feels much calmer and less "rising the mob", instead something aimed at trying to help users, and showing what the boards have been doing on older bios versions.

I dont have a strong opinion right now on recalls, class actions and so forth, Intel probably need a chance to see if they can at least stabilise chips that havent degraded to the point of no return, alongside RMA'ing faulty chips, it is good they extended the warranty, something they should have announced at least a few weeks back.

 

[f0ssile]

More than calm it seems fake, and for a change it shifts too much responsibility towards the MBs.

We will see the altars in the reviews of the new Ryzen...
Will they pretend nothing happened by setting the set above Baseline?
Why is there no default set? (the story stinks too much)
Will there be mixes with base and extreme?
Is Extreme the OC of Intel, or does it want to seem like the default?

Why does the guy avoid what Intel has seriously done wrong?
Where are the nonsense specifications given by Intel, analyzed by Steve?
When he put the 14900K, what did he mean that Intel is more stable? (well, we saw...)
Why did he make a video 5 days before in which he pretended to know even less?

 

[TumbleGeorge]

intel is denying us warranty

Blue/green slave detected.

 

[mkppo]

I'm hell bent on providing facts. Those are the 3 popular cases with threads / videos on youtube and reddit about Intel "denying RMA". They are spammed everywhere, even though none of them have intel actually denying RMA.

But sure, if you prefer me to go along with the intel bashing sure, let's do it

What good is a 5 year warranty when Intel denies 100% of the RMAs, multiple users are complaining on reddit left right and center. Terrible company, ill only buy amd from now on.

Who told you that prebuilds don't have or have 1 year warranty? Maybe it's a US problem? Cause in EU it's illegal, you can't even sell a product with only 1 year of warranty. 2 years is the minimum.

There are quite a few things you're conveniently omitting from those cases, but let's ignore that. Try to understand that there will be an uptick in returns for higher end 13/14 gen parts and there will always be users that try to benefit from this scenario. There are also legitimate cases of many server farms not given warranty and users having to go through hoops, i suggest you look up or watch some videos because there's a lot of evidence out there. It's normal for companies to mess RMA up, you can't be perfect so why wouldn't there be. It's just being made to be a bigger deal than usual when the RMA dept messes up

If prebuilts come with two year warranty, my questions still stand. Because it's been two years since launch, and most are out of warranty.

 

[JustBenching]

There are quite a few things you're conveniently omitting from those cases, but let's ignore that.

No, let's not. If im omitting anything important it's not on purpose and i geiniunely want you to correct me. After all RMA coverage is really important, show me which of those 3 cases I was wrong on and ill be on the same side as you exposing intel for denying RMAs.

If prebuilts come with two year warranty, my questions still stand. Because it's been two years since launch, and most are out of warranty.

What do you do when any product is out of warranty and it fails? I don't get what is the point of the question, you could have asked the same about any product in the history of products. Had a TV fail just outside warranty with minimal usage, like 4 hours a week. Happens.

 

[mkppo]

No, let's not. If im omitting anything important it's not on purpose and i geiniunely want you to correct me. After all RMA coverage is really important, show me which of those 3 cases I was wrong on and ill be on the same side as you exposing intel for denying RMAs.

What do you do when any product is out of warranty and it fails? I don't get what is the point of the question, you could have asked the same about any product in the history of products. Had a TV fail just outside warranty with minimal usage, like 4 hours a week. Happens.

I can't be bothered to actually search and write up on a case by case basis. People can do that in their own time but there's going to be a more cases as the weeks go on..

I don't know how I can make the question any easier, I thought it's simple. These consumers (which represent the majority) don't get the 3 year standard CPU warranty nor the extended warranty which intel initially stated on their website they'll cover but later removed. Instead of adding two years and making it a big deal, including tray CPU's would've been better for everyone and I asked if you'd agree.

edit: oh and i mentioned some cases you ignore, like a few devs and server farms detailing their issues with Intel RMA. Do you honestly think they haven't messed up RMA at all? Strange.

 

[JustBenching]

I can't be bothered to actually search and write up on a case by case basis. People can do that in their own time but there's going to be a more cases as the weeks go on..

So why did you say im omitting stuff? Im not, those are literally the 3 most popular cases that "intel denied rma" on reddit and forums. I covered all 3 of them.

I don't know how I can make the question any easier, I thought it's simple. These consumers (which represent the majority) don't get the 3 year standard CPU warranty nor the extended warranty which intel initially stated on their website they'll cover but later removed. Instead of adding two years and making it a big deal, including tray CPU's would've been better for everyone and I asked if you'd agree.

Yes, including 2 years for everyone would have made it better. Heck, including 20 years for everyone would make it even better. While we are at it, we can make it 50. Im not here to argue what would have been better. I'm saying that what applies to any other product applies to intel products as well. If your product dies after the warranty expires, well, you have to buy a new one.

But with all that said, intel announced that a similar extension is on the works for tray cpus as well. Not that that will appease anyone of course, people will still be complaining on this and other forums. Theyll find a reason, im sure.

edit: oh and i mentioned some cases you ignore, like a few devs and server farms detailing their issues with Intel RMA. Do you honestly think they haven't messed up RMA at all? Strange.

And there are also many cases that intel exchanged multiple (and I really mean MULTIPLE) chips for server farms. These are all the definition of anecdotes. If you want to be more specific and send me something more particular id be glad.

 

[mkppo]

So why did you say im omitting stuff? Im not, those are literally the 3 most popular cases that "intel denied rma" on reddit and forums. I covered all 3 of them.

Yes, including 2 years for everyone would have made it better. Heck, including 20 years for everyone would make it even better. While we are at it, we can make it 50. Im not here to argue what would have been better. I'm saying that what applies to any other product applies to intel products as well. If your product dies after the warranty expires, well, you have to buy a new one.

But with all that said, intel announced that a similar extension is on the works for tray cpus as well. Not that that will appease anyone of course, people will still be complaining on this and other forums. Theyll find a reason, im sure.

And there are also many cases that intel exchanged multiple (and I really mean MULTIPLE) chips for server farms. These are all the definition of anecdotes. If you want to be more specific and send me something more particular id be glad.

I said you're omitting stuff because you are. Look, no one has the time to summarize every case of denied RMA and this is really wasting everyone's time. Also, no idea why you are going on a 20 year old tangent, there's literally no point you're making there. I'll quote one, there's a post by jerubedo in reddit two days ago (which i think is your first point?) outlining intel identifying both his CPU's as tray CPU's (which Intel should cover anyway but decided not to) as opposed to boxed even though he provided serial numbers, pictures and all that and were bought from Microcenter and Amazon so they definitely weren't tray. Then, after Intel identified these incorrectly as tray CPU's followed up with this quote: "However, if the products fail the validation process, the units will be retained and confiscated, and no replacements or refunds will be provided"

There are a few issues. First, most retailers will not offer refunds in the US months after purchase. He got lucky, not everyone is. And you have to understand that refunds are very much something one might want in this case. Then there's the case of Intel being incorrect and has nothing to do with thermal paste hiding the serial number like you quote. He mentioned the TIM being on the side and is absolutely not a case of denying RMA. Intel also quoted incorrect SN's for his CPU's and a whole heap of stuff happened which I won't get into but you get the picture. The process was a whole heap of mess and the retailers bailed him out (when they weren't obliged to).

The problem is, you are not the designated RMA defender for Intel. There WILL be failed RMA cases, it would happen to anyone and seems very much the case that Intel messed it up here and will mess up in the future as well. Does that mean they are denying it to everyone? No. But you cant individually defend every single case and say "oh no Intel was right". In many cases, they weren't. And that's fine.

Your argument of 'what applies to any other product applies to Intel as well'? If that's your stance, sure and we'll agree to disagree, because others aren't designing chips that are failing. I think otherwise; if you've designed a product that lasts much less than what it should, you shouldn't just put out a relatively useless public statement of warranty extension that doesn't affect most of their users. You should disclose batch numbers affected by oxidisation, you should put out a tool that actually tests for stability (Intel's tool is laughably useless as it shows CPU's are stable when they're not) and whenever that test fails, there should be an RMA issued regardless of whether the CPU is tray or boxed. There's other stuff they can do but these at the bare minimum would make it right. The two year extension does next to nothing.