Intel Needs to Say Something: Oxidation Claims, New Microcode, & Benchmark Challenges

We don't know what the cause of failure is yet. We've received a lot of leaks and credible tips and, working alongside Wendell of Level1 Techs, we have been trying to get to the bottom of Intel's CPU failures. The current trouble is how reviewers should even test these parts for the upcoming AMD Zen 5 reviews though, as any changes from Intel will need to be reflected in that comparative data. Likewise, until Intel puts out a better statement, we simply do not feel comfortable recommending 13th and 14th Gen CPUs while this matter is pending. This video talks about tips we've received relating to oxidation or potential corrosion within the silicon, specifically relating to the vias, and other tips like excessive voltage, memory speeds, unstable frequency, and failure rates. This is not a complete investigation. We are still working and do not yet have a conclusion, nor do we have full confidence in any one failure mode.

You can also email tips to us at our tips line.



TIMESTAMPS

00:00 - Why We Can't Recommend Intel Right Now
04:57 - HELP US RESEARCH!
05:54 - The Story So Far (Recap)
08:47 - New Information
11:03 - Current Claims & Tips
15:33 - Oxidation
19:32 - Pending Failure Analysis Results
20:53 - Failure Rates
24:49 - Important Reminders
26:20 - What We Do as Reviewers

** Please like, comment, and subscribe for more! **

Links to Amazon and Newegg are typically monetized on our channel (affiliate links) and may return a commission of sales to us from the retailer. This is unrelated to the product manufacturer. Any advertisements or sponsorships are disclosed within the video ("this video is brought to you by") and above the fold in the description. We do not ever produce paid content or "sponsored content" (meaning that the content is our idea and is not funded externally aside from whatever ad placement is in the beginning) and we do not ever charge manufacturers for coverage.

Follow us in these locations for more gaming and hardware updates:


Steve Burke: Host, Writing, Video Editing
Vitalii Makhnovets: Video Editing

[keggerous]

Not knowing whether you're going to have a working CPU in the future and not knowing how the company that sold it to you will be handling the issue is INCREDIBLY frustrating.

[Erelyes]

$5 says AMD was about to release pricing for the 9000-series and then they saw this going down - and went "well, wait just a minute now"

[orion]

Oxidized copper on the inside of the CPU? Well that's one way for Intel to go green.

[dansanger]

I can't imagine that someone who set out to start a YouTube channel for gamer enthusiasts thought that there would be this much investigative journalism involved.

[MastrCake]

“Some of the examples of instability include…BSODs…”

Oh, so Crowdstrike compiled today’s update on an Intel CPU. Got it.

[VoidsentVivi]

One of my more technical inclined friends almost went nuts trying to pin down the issue with his 13900k. The crashes he was experiencing were so sporatic. Turns out one of his cores had died and once he disabled it, the crashes stopped.

[BeardedFrog]

My main thought here is how Intel is even going to resolve this. No matter WHAT the culprit/root cause is, there is no easy or especially cheap way Intel can remediate this.

If they push "fixes" that lower frequency, memory support, or essentially anything that changes the advertised design specs, it's an extreme case of false advertising for any customer who purchased the chips prior to the change (not to mention it further lengthens the gap of AMD's leads).

You can't buy a V8 car with 500hp, that has global issues with misfiring suddenly, and the "fix" is to deactivate two cylinders. Thus making it functionally a v6 with significantly less power. It doesn't work that way. People paid for a product as advertised, and may have made a different choice if the product had these different specifications. This is not a viable solution, and would not only tarnish their reputation further, but would be a massive legal issue.

The only ACTUAL option they have (again no matter what the culprit actually is), assuming they can't push a "fix" that does not alter advertised specs, is a recall.

This would be extremely costly as this issue seems to have permeated all matkets across many product lines, over a lengthy but unknown amount of time. Recalling all these CPUs would cost them an insane amount of money, and would also still tarnish their reputation significantly even if done as efficiently as possible. This assumes that the culprit is something that _can_ be resolved (aka manufacturing issue, rather than inherent design flaw).

Their only true hope here to resolving this issue is that they can release a "fix" via microcode update or otherwise to customers that in no way alters the advertised design specifications/performance of their processors. Their reputation will still take a hit, but it will be the only viable "cheap" way they can pull themselves out of this situation. I suspect their silence is because they are running around like chickens with their heads cut off trying to find a way to do just that to avoid the significantly more costly outcomes.

While I have personally found Intel to be a completely illogical choice compared to AMD since at least AM5 release, and find no redeeming qualities of current Intel CPUs; as a consumer, I sincerely hope they get their act together and can resolve this properly and compete properly again. Otherwise if AMD continues its dominance, we will eventually see a reverse of the prior times in history where Intel was in such a position with minimal competiton. This was a time when the CPU market/prices prices were like the current GPU market/prices and the top Intel CPUs were like £1k.

[auxityne]

I'm so old, I remember when "Intel Inside" was a fun jingle, not a death threat.

[PlayNeth]

Intel's GPU department feels like a whole different company at this point.

[philh]

So its interesting you say that. I work for a large government organization. I am always skeptical of BSODs in enterprise environments as they install all sorts of heavy SEIM and productivity suites on their machines. Midway through last year, our government program saw an elevated rate of GFE (laptop) BSODs that has plagued new team members regularly. Often, GFE reimages do not resolve the issue. Again, the I'd say about 20% of new team member's GFEs are affected. Virtually all the GFE's are $7000+ Dell enterprise Intel laptops. I'll try to find what processors they are using. I know this is only focusing on desktop processors, but I'm starting to wonder if this affects mobile cpu skus as well.

[devonmoreau]

if this is a manufacturing defect, this is extremely concerning for the reputation of Intel foundries

[EkiToji]

So I'm not in the industry but did years of grad school work with atomic layer deposition (ALD.) I've not personally worked on plasma-enhanced ALD (PEALD) but you really shouldn't have any oxygen anywhere between the metal-organic or metal-halide tantalum precursor and whatever they use for the nitrogen, which I'd be surprised if it's anything other than ammonia or an N2 plasma. The whole damascene process used for copper interconnects is very old and well studied so they'd have to really screw up their process flow, although ALD hasn't always been used but it's just great because of being able to cover high aspect ratio structures.

It's important to realize too that copper can't touch the silicon directly or it will diffuse into it and form deep-level traps that will absolutely ruin your performance. You instead rely on some other thin metallization steps and form silicides with those metals to tune your metal-semiconductor junction and the Schottky barrier height. You can look up the Schottky-Mott rule for more on that. In short though any of that stuff would have been caught well before production, so I doubt ALD and the various process steps for those interconnects are to blame. It would be interesting to see if there are different results from these CPUs depending on if they were manufactured in Hillsboro, OR or Chandler, AZ however.

If oxidation is a problem it's definitely a very embarrassing issue considering how hard that would be to have happen. My worry would be it's a defect on one or more lithography masks if that's the case but for it to show up so slowly is just odd considering how hard of a diffusion path it would have. In any case I'm a couple decades removed from the field but if I can help as a sanity check or someone to confirm somewhat basic technical details feel free to hit me up.

Also I really don't envy the FA guys you hired because it will be a literal needle in a billion haystacks if it's not one of the giant interconnects. People seem to think it might be related to the ring interconnect as that was a major change from 12th to 13th gen so they may be able to find something without sinking way too much time into it.

[mikoldeon]

PC users are at the moment, where they can't be sure their (usually expensive) Intel CPUs are stable at _default_ settings... it's insane.

[BlackHawk]

I remember my 2500K going strong for years and years even on OC
How the mighty have fallen

[ShamanicEnzan]

Can’t wait for the UserBenchmark spin on this.

“Intel cpus are too powerful not even god can contain their strength”

[Unpluggedx]

As someone who has a 13700K and is experiencing several crashes per day I thank you for being relentless to get answers for this. You guys are heroes!

[circuitsoldiers]

I'm on my second i9-13900K after the first one degraded rapidly. Heading to the survey rn. Thank you so much GN ❤

[RILDIGITAL]

The potential impact on the tech industry and Intel's reputation could be significant. Maintaining transparency and speedily addressing these issues should be Intel's top priority.

[uncleanriches]

I have a 10900K and nearly upgraded it, my housemate has a 13700K and it's his second. The first one became temperamental and there were lots of unexplained system crashes while gaming. He like anyone else did not suspect the CPU and we benched swaps of all other likely components to cause the issue and it was not resolved. He RMA'd his 13700K and Intel reported that it was faulty and sent him a new one, the new CPU ran fine for a few weeks but the crashing has begun to return.

This new information has been a real eye opener!

[maze]

i love how alot of other hw news outlets just said "100% failure rate on all intel cpus" instead of the correct version "100% of failures at some server providers are intel cpus"