News Intel announces an extra two years of warranty for its chips amid crashing and instability issues — longer warranty applies to 13th- and 14th-Gen C...

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
Been on my mind since the start of this fiasco. Those so-called normies should not have been getting anywhere near i9k/ks CPUs and dropping them into enthusiast MBs hitting the power switch and walking away.
They were never the intended market, intel has always inferred whether written or otherwise, that they were for people who knew their way around a modern bios and also would know what would be bad long term, like giving the preferred cores all the beans, and running them on cooling not up to that level of heat removal.
I completely agree with this but the fault lies with intel. They are not really promoting their non k chips, they arent even sending samples for review. I don't get why, since theyd easily be winning the highest efficiency if they just sent a single non k sample and everyone (hopefully??) would stop talking about how far behind they are in efficiency when in fact they are leading out of the box, and not by a small margin. I really don't get it.
 
Last edited:
Intel to trade commission oh look we did something for the consumers. Also aren't these still being sold? Lamen consumer oh look these Intel cpu have 2 extra years of warranty compared to Ryzen. 🙃
 
Other than the obvious worry I now have about my CPU dying on me.
I'm also worried about overclocking as a whole.
Processors are getting hotter and more power hungry. GPUs too I guess.
I hope we never reach a day where we're unable to push parts even a little bit past their spec without worrying that they'll break down. But it kinda seems like we're on the way there.

I also hope being on an 'enthusiast-level' board will help. I think I heard it said somewhere that higher quality boards are less likely to run into issues, but I can't remember where and I'm not sure that it's even true. Maybe wishful thinking.
 
Not good enough. In fact: not good at all.

Comparing this to my i7 2700K which I had running for 10 years at 4.6Ghz with a day 1 overclock (I even used Tom's guide from back then), it's just not acceptable.

CPUs, as any other electronic component, does not degrade in any meaningful way during the life span of a human being when it is working "within spec", as the whole dang reason the "spec" is written is for the CPUs to work for hundreds of years under those conditions if needed. The fact these CPUs are failing "within spec" is an abomination.

The recall is the only good solution for people who will not know about this issue and for the Tech Press to put a big disclaimer everywhere when talking 13th and 14th gen so they don't buy a used one without knowing. Remember Intel CPUs do keep a good value over time. Small distributors will have a big pain as well with this. If Intel is extending the warranty, then they have to do as well. This is not a simple financial hit they can take and Intel is doing them absolutely dirty with this.

Think outside of your comfort DYI'er zone and how Intel's inactions/missteps affect the whole chain.

Regards.
 
  • Like
Reactions: Flayed
Not good enough. In fact: not good at all.

Comparing this to my i7 2700K which I had running for 10 years at 4.6Ghz with a day 1 overclock (I even used Tom's guide from back then), it's just not acceptable.

CPUs, as any other electronic component, does not degrade in any meaningful way during the life span of a human being when it is working "within spec", as the whole dang reason the "spec" is written is for the CPUs to work for hundreds of years under those conditions if needed. The fact these CPUs are failing "within spec" is an abomination.

The recall is the only good solution for people who will not know about this issue and for the Tech Press to put a big disclaimer everywhere when talking 13th and 14th gen so they don't buy a used one without knowing. Remember Intel CPUs do keep a good value over time. Small distributors will have a big pain as well with this. If Intel is extending the warranty, then they have to do as well. This is not a simple financial hit they can take and Intel is doing them absolutely dirty with this.

Think outside of your comfort DYI'er zone and how Intel's inactions/missteps affect the whole chain.

Regards.
If you are referring to OEMs then that responsibility falls on them, whether it’s a laptop or PC. Intel has no information on OEM purchases.
 
Other than the obvious worry I now have about my CPU dying on me.
I'm also worried about overclocking as a whole.
Processors are getting hotter and more power hungry. GPUs too I guess.
I hope we never reach a day where we're unable to push parts even a little bit past their spec without worrying that they'll break down. But it kinda seems like we're on the way there.

I also hope being on an 'enthusiast-level' board will help. I think I heard it said somewhere that higher quality boards are less likely to run into issues, but I can't remember where and I'm not sure that it's even true. Maybe wishful thinking.
I think we are on cusp of not needing a OC, out of the box we are already close to peak, all CPUs bins are different, however, board partners are already pushing the out of box limits.

Maybe x86 has reached its EOL.
 
If you are referring to OEMs then that responsibility falls on them, whether it’s a laptop or PC. Intel has no information on OEM purchases.
OEMs like Dell and HP are covered by different types of agreements. Not sure about Asus, but I'd imagine they can absorb the initial cost of this with not much problem.

Think about smaller SIs (System Integrators) and your classic "mom&pop" shop that also builds PCs. They're the ones left with holding the bag for Intel with no compensation (that we know of, that is). Looking at how Intel is cutting a lot of costs left and right, I doubt they'd even care about medium and small businesses.

Regards.
 
For boxed processors? Intel isn't going to pay you for the time it takes to swap your CPU and for loss of uptime. That's not part of the warranty you get with boxed retail CPUs.

As for OEMs and system integrators that are buying by the tray, I used to work in a supply chain where my company supplied raw castings to another company that would machine them and sell them on. We had an agreement where below a certain reject rate they would just return the parts for investigation and get replacements, but if the reject rate went above that we'd owe a penalty for the lost throughput. I'd assume there's a similar agreement between OEMs and Intel, where the OEMs bear the labor cost of warranty work in return for cheaper OEM prices up to some point, but Intel owes some kind of rebate/discount on future orders if the warranty rate climbs into atypical territory.
What is the reject rate on 13th and 14th gen Intel chips though? It seems that many are assuming and stating 100% when it may be low enough per the agreement your old company had to just offer replacements.
How bad this is depends on the rate. If it is less than ram, or comparable to previous Ryzen generations then it isn't that big a deal. If it is really high then it is a big deal.
Yeah, I seem to recall reading that the CPU burns some fuse or something, if it's ever overclocked. Intel explicitly says overclocking voids the warranty, so you'd expect them to embed a mechanism so they could tell if it happened.

Something like undervolting is probably a different story, though.
I've only heard of that with AMD chips but it wouldn't surprise me if Intel did it as well.
Other than the obvious worry I now have about my CPU dying on me.
I'm also worried about overclocking as a whole.
Processors are getting hotter and more power hungry. GPUs too I guess.
I hope we never reach a day where we're unable to push parts even a little bit past their spec without worrying that they'll break down. But it kinda seems like we're on the way there.

I also hope being on an 'enthusiast-level' board will help. I think I heard it said somewhere that higher quality boards are less likely to run into issues, but I can't remember where and I'm not sure that it's even true. Maybe wishful thinking.
If you are worried then you can alleviate that worry by doing something about it.
Since it is probable that a significant portion of the instability is due to voltage induced degradation then keeping your volts below what would degrade them would be the answer. If you can, adjust your bios settings to keep your volts below 1.4v, even spikes. That goes for Ryzen as well. And basically all CPUs back to 2010. Play around a bit with undervolting. You may find your chip needs significantly less than review power.
 
  • Like
Reactions: slightnitpick
I also hope being on an 'enthusiast-level' board will help. I think I heard it said somewhere that higher quality boards are less likely to run into issues, but I can't remember where and I'm not sure that it's even true. Maybe wishful thinking.
Perhaps, but only perhaps. Enthusiast-level boards often have features and overkill-solutions to somewhat non-existing problems, but I am not sure they're manufactured to a higher level of quality. Maybe the most obvious is higher grade capacitors. The quality level of midrange boards (manufacturing-wise) is astonishing with extremely low failure rate.
 
I'm also worried about overclocking as a whole.
Processors are getting hotter and more power hungry. GPUs too I guess.
I hope we never reach a day where we're unable to push parts even a little bit past their spec without worrying that they'll break down. But it kinda seems like we're on the way there.
As long as the "silicon lottery" exists, I think overclocking will be here to stay. My rationale is that AMD caught flak for advertising boost speeds in Ryzen 5000 that many CPUs would rarely ever reach, or at least sustain for more than a few milliseconds at a time. So, they learned they had to be more conservative in what boost speeds they'd advertise, which means incorporating bigger margins. That means that a top-quality part should still have room to stretch its legs, over and above the specified frequencies.

Granted, we're probably talking about only a few hundred MHz, but it's still something.

Maybe x86 has reached its EOL.
According to the lead architect Zen:

"... a feature that looks great in 3nm not looking so great in 4nm because of the power impact of the not-as-efficient transistor and how it affects the floorplan. Normally, we do the architecture in one, and then we port on the next one, and then you have a lot of time to deal in the floor plan with the two technologies. [..] It was just really challenging. But that gives Zen 6 a lot of room to improve."

Source: https://www.tomshardware.com/pc-com...ast-3nm-launch-zen-5c-cores-for-desktop-chips

Sounds exciting!
 
Last edited:
This doesn't matter to over clockers, as noted on the Intel site...
"Altering clock frequency or voltage may void any product warranties and reduce stability, security, performance, and life of the processor and other components. Check with system and component manufacturers for details."
Only normies that put up with out of the box configurations is this news relevant too.
I got a 13900k at the end of 2022, and overclocked it to the hairy edge of stability. That said, I manually set the voltage at 1.35 and even with extreme llc profile I've never seen excursions past 1.4. I've been running that chip pretty hard all this time, and I haven't seen any undesirable behavior.

Sample size of 1 isn't worth much, but it has me wondering if the overclockers are in better shape vs those that accept the default Intel/motherboard settings.
 
Warranty and life expectancy are not the same. Do you expect all your belongings to break when they reach the end of their 1 year warranty?
I never said I expect them to break. I'm saying the company expects a big percentage of them to actually do break, that's why it's not financially viable or lucrative to give you a bigger warranty.

Say a psu having a 10 year warranty. It doesn't mean that the company is suddenly a saint and wants to replace your failing psu, it just means they are freaking sure that their product is so high quality they won't need to.
 
  • Like
Reactions: helper800
The symptoms of overclocking being strikingly similar to the symptoms of the failing CPU's in question....they'd better just process the RMA and stop talking. Intel and AMD both have a long history of honouring warranties on even admittedly overclocked chips (PCWorld had an article on this iirc). This is either just good faith, or they cannot tell/did not care. Who knows? Either way, I'm pretty sure Intel knows they've stepped in it now, I don't expect many people to have any issues with the RMA process anymore.
This.

There is a class action lawsuit already happening. Intel knows that it needs to honor RMA's because that is the only way they can prevent the suit from going to court, which all companies will want to avoid.
 
  • Like
Reactions: slightnitpick
Thanks for pointing this out. The same can be said for their Best CPU for Gaming in 2024 article.

Both articles have a paragraph, near the beginning, cautioning about Intel CPUs. However, probably 90% of readers will just skip down to the actual recommendations, and the editor didn't bother to list degradation or the possible performance impact of the microcode update as a Con, on any of the three Raptor Lake (B0) K-series CPUs included in the recommendations.

Sadly, the comments of that article have been disabled for the past 5 years. They just keep updating the same article and nobody bothers either to unlock the comment thread or create a new one for it.
It's a fair point that the issue is not listed in the breakouts for each chip in the list of the Best CPUs for Gaming—I've added that.

This has been a long-developing situation, so we are attempting to approach it as well as we can. Chip issues are common—AMD has had chips that don't meet boost specs, chips that have melted both the processor and the socket, and Nvidia has a series of GPUs that continue to have melting problems with the power connectors.

Unfortunately, we do not know how prevalent those issues are — do they impact 0.2%, 0.5%, or 10% of the products sold? That makes it exceedingly hard to make blanket statements about purchasing decisions based on those reports. As such, we haven't attached warnings to the AMD and Nvidia products, either during those issues (AMD), or on an ongoing basis (Nvidia), for instance.

By our own decision, we already added warnings to the guides for Intel processors as the scope of Intel's issue has become more known. This is also spurred by confirmations that Intel issued last week. We also just learned this weekend that the problem extends to 65W processors, so this is a fast-developing situation.

In either case, for the Best Gaming CPUs, we already recommended only two lower-end Intel processors as primary picks; the other four (high-performance) were all AMD. The 13600K is the lone impacted processor that we recommended, but only because we did not know that the issue impacted 65W chips at the time. That pick has now been downgraded.

For the AMD vs. Intel article, I updated this and Best CPUs recently to add the warning. As you can see, the AMD vs. Intel article is very expansive with multiple categories, and it needs an overhaul — a time-intensive process. This was slated for completion after the Zen 5 launch, when we have a fresh batch of benchmarks and results to rework the article and also when we can assess the impact of the microcode patch. The worry is that an incomplete update just becomes a huge discombobulated mess, but I had already reworked the introduction and the conclusion to reflect the impact of the instability. However, I just updated the article further to drive home the impact of the instability issues. So now it's a mess, but so it goes.
 
Last edited:
Up until my current build I've always had Intel processors, but the cost/performance wasn't there for the then new 13th-Gen, leading me to go to AMD for the first time ever. Seeing all this, I'd be *very* hesitant to go back to Intel.
 
  • Like
Reactions: -Fran-
Thats good news, but they need to honor it. They were already rejecting the RMAs with existing warranties.
...
It is good news but indeed feels like more of a publicity stunt and late damage control action if they did indeed reject tens, hundreds, or more of RMA claims based on this issue, which is why the class action lawsuit prospect has started and discovery is under way.
I can only hope the tool that they are making is fair and accurate and greatly assists in approving and moving the RMA process forward for individuals and organizations.
 
I completely agree with this but the fault lies with intel. They are not really promoting their non k chips, they arent even sending samples for review. I don't get why, since theyd easily be winning the highest efficiency if they just sent a single non k sample and everyone (hopefully??) would stop talking about how far behind they are in efficiency when in fact they are leading out of the box, and not by a small margin. I really don't get it.
Because their boost power consumption isn't really that low, and it more than likely can't show significant performance especially in single thread workloads or games etc. to claim they are the best..

It is good news but indeed feels like more of a publicity stunt and late damage control action if they did indeed reject tens, hundreds, or more of RMA claims based on this issue, which is why the class action lawsuit prospect has started and discovery is under way.
I can only hope the tool that they are making is fair and accurate and greatly assists in approving and moving the RMA process forward for individuals and organizations.
It likely will be a death sprial to Intel, say in 3 years anyone got unstable CPU, be it really OC to death or the degradation get denied RMA, this will haunt them back, also it doesn't help that the RPL is literally the LAST Core i9 series in history and that it will be a stinky last...
 
Because their boost power consumption isn't really that low, and it more than likely can't show significant performance especially in single thread workloads or games etc. to claim they are the best..
Boost power cosunmption is 106w (for 54 seconds) for T and 219w for the non k. Both way lower than the competition, even for those 54 seconds.

Single threaded workloads


Seems like they beat the competition in both ST and games, even capped as low at 95 or 65w.
 
I never said I expect them to break. I'm saying the company expects a big percentage of them to actually do break, that's why it's not financially viable or lucrative to give you a bigger warranty.

Say a psu having a 10 year warranty. It doesn't mean that the company is suddenly a saint and wants to replace your failing psu, it just means they are freaking sure that their product is so high quality they won't need to.
I would guess that the primary issue with a greater than three year warranty for processors is that Intel/AMD don't make the older processors that long. So if a person gets a warranty replacement after 3 years it will be a new processor of another generation. This will likely require that the person buys a new motherboard, which isn't going to result in a happy customer. So Intel/AMD end up shelling out for a new processor and still have an unhappy customer.

OTOH, it's relatively straightforward to replace one company's PSU with another company's PSU. A 10 year warranty is part of the sales pitch.

On top of that, most companies and individuals upgrade more often than every 10 years. Intel/AMD are incentivized to encourage this upgrading (disposable culture). A longer warranty is a lose-lose situation.

I do wonder if they tested the warranty period with marketing surveys. It wouldn't surprise me.

And if warranties for processors did last 10 years, and did result in a newer generation processor getting shipped as a replacement, there'd be guides online about how to screw up your 9 year old processor in just the right way to get a free, modern replacement. Even if it just resulted in a refund there would still be these guides online. And getting a replacement every 9 or 10 years wouldn't necessarily trigger a warranty claim rejection, but getting one every 2 or 3 years would start to trigger this rejection.
 
I never said I expect them to break. I'm saying the company expects a big percentage of them to actually do break, that's why it's not financially viable or lucrative to give you a bigger warranty.
It's not that they necessarily know they'll break, but there's always a risk the failure rate will tick up higher than they expect - this Raptor Lake issue being the perfect example. Except, in the case of Raptor Lake, the issue is so bad that Intel had to step up to save its reputation. But, if it had been less severe, then maybe they wouldn't have offered the warranty extension and would be really glad the warranty was only 3 years.
 
  • Like
Reactions: helper800
It's a fair point that the issue is not listed in the breakouts for each chip in the list of the Best CPUs for Gaming—I've added that.
Thanks, Paul! I would really appreciate it, if you could re-test Raptor Lake with Intel's microcode patch, once it's released. Please also adhere to their official guidelines for what settings people should use to avoid premature CPU failure.

In either case, for the Best Gaming CPUs, we already recommended only two lower-end Intel processors as primary picks; the other four (high-performance) were all AMD. The 13600K is the lone impacted processor that we recommended, but only because we did not know that the issue impacted 65W chips at the time. That pick has now been downgraded.
I'm also seeing:

Intel Core i7-14700K
Overall Best CPU for Gaming — Alternate Pick


Intel Core i9-13900K
Highest Performance Best CPU for Gaming - Alternate Pick

All three have the caveat, so I guess you indeed noticed them.
 
I have to wonder how difficult the warranty process will be. I've already had MASSIVE issues in trying to deal with Intel and their absolutely stupid outsourced tech/warranty support this year already. Now I imagine that, along with them also insisting on purchase receipts to make sure the CPU hadn't been resold.
 
Status
Not open for further replies.