Mr Intel leaving Intel is not a great sign... for Intel

Interesting column in The Register today.

https://www.theregister.com/2024/12/09/opinion_colum...

I'm not sure where this will all end. One thing I think however is certain; the mighty and once unassailable Intel some of us may remember from the halcyon years of the PC revolution from the 1980's through 2000's is well and truly gone.

In reply to a post by Pheasant:

One thing I think however is certain; the mighty and once unassailable Intel some of us may remember from the halcyon years of the PC revolution from the 1980's through 2000's is well and truly gone.

Thanks for link, I’ve been listening to various podcasts, and even the Apple centric ones think that Gelsinger going leaves intel with a CEO hole that plugging is going to be really hard. So I agree with your statement above.

Rupert Goodwins has been negatively writing about IT industry for decades. I first came across his output in Personal Computer World magazine… along with Dick Pountain whom is now at PCPro.

Not sure Rupert has the Intel and Apple stuff right, as the Macbook Air relied on the Pentium M derived Core 1 (duo, not the brain dead solo), and then the Core 2,3, etc and Apple’s 11” Mac used the entry level CPUs for ages. Dramatically lower battery consumption than the PowerPC architecture from IBM/Motorola.

The future being ARM is good for laptops, lets see if the ARM world kicks the data centre and enterprise compute.

Edited by jchamier (Mon 09-Dec-24 18:13:26)

In reply to a post by jchamier:

Not sure Rupert has the Intel and Apple stuff right, as the Macbook Air relied on the Pentium M derived Core 1 (duo, not the brain dead solo), and then the Core 2,3, etc and Apple’s 11” Mac used the entry level CPUs for ages. Dramatically lower battery consumption than the PowerPC architecture from IBM/Motorola.

In reply to a post by jchamier:

The future being ARM is good for laptops, lets see if the ARM world kicks the data centre and enterprise compute.

Amd were, under the plug in play bulldozer arch, had a plug in arm unit. Plus there was talk of a dedicated opteron arm chip - which i think got burried either when Lisa came along or just at the end of Rory's tenure

i still can't understand why may sold arm off to softbank

Not read the article. But i do think Intel has said and done one too many lies, stifled the x86 community for fair to long and AMD's third cometh seems to be stable this time ..

Add in Intels multiple chip disasters - will Intel ever be the same?

Unless you have support structures that rely on intel and or software than runs better on intel, then why buy intel!

It’s poor. Really poor from the Intel board. Smacks of desperation.

Have a read if you have 10 minutes. There’s nothing there that one could really argue with, despite being slightly hyperbolic; it is after all an opinion piece. I certainly wouldn’t argue with the conclusions, even if some of the details aren’t perfect.

Intel have been running on “milk x86 autopilot” for way too long (decades!!) and it’s all finally caught with them.

I fully expect them to break up, piecemeal like going forward. A sad slow decline.

More today from The Register.

With Gelsinger gone, to fab or not to fab is the $7B question

Quite the pickle.

Edited by Pheasant (Tue 10-Dec-24 22:41:37)

Link doesn't work, should be https://www.theregister.com/2024/12/10/gelsinger_dep...

Edited by billford (Tue 10-Dec-24 22:44:45)

Thanks. Fixed.

For a few years now I've often been reminded of the early mickey-taking AMD ads- a coffin and headstone with "Intel inside" on it... a bit premature but they might have it right this time.

Yeah I remember the ads too. Cheeky but perhaps a touch prophetic. Saying that AMD and Intel are besties now like they've never been before...

Hell Freezes Over as AMD and Intel Come Together for x86

Maybe AMD are just waiting for the right moment to....eat the elephant, or what remains of it....

It might well be that intel's name goes.

That said, i think group is there from the threat of google, amazon and nvidia all producing their own chips(cpus), even Microsoft has got into the act

AMD would be the logical choice to swallow the x86 / product side of the business. Perhaps with the AMD-Intel brand in tow, it would make big corporate customers feel all warm and fuzzy, knowing that there was "Intel inside" even if only skin cardboard deep.

The Intel foundry business is bleeding cash. Just maybe TSMC or Samsung would buy it. But successive failures developing future processes would make it less than appealing. TSMC already make a whole bunch of stuff under contract for Intel including their latest Arrow Lake and Lunar Lake CPU components and they are miles ahead of Intel in foundry capability.

In reply to a post by Pheasant:

AMD would be the logical choice to swallow the x86 / product side of the business. Perhaps with the AMD-Intel brand in tow, it would make big corporate customers feel all warm and fuzzy, knowing that there was "Intel inside" even if only skin cardboard deep.

It would at this juncture make sense to do. Even with Jim keller's help to bring some amd "glue" intel are still in a mess. Intel still though makes a ton of cash in other areas though. They completely messed up Optane for instance, charged royalties for Thunderbolt etc and remember the physical keys you had to buy to unlook cpu features.

In reply to a post by Pheasant:

The Intel foundry business is bleeding cash. Just maybe TSMC or Samsung would buy it. But successive failures developing future processes would make it less than appealing. TSMC already make a whole bunch of stuff under contract for Intel including their latest Arrow Lake and Lunar Lake CPU components and they are miles ahead of Intel in foundry capability.

Reminds me of AMD's foundry, back then they were several nodes behind and couldn't execute and intel was at that time executing well - before 14+++++++++++ episode. Even now Glo Flo can't do what TSMC (i always say tmsc in my head for some reason) can do in terms of nodes. At least amd didn't lie unlike Intel's stupid stupid comments

I think Rory sold the foundry arm off - i can't remember but if he did that was in hindsight the best thing amd had done. Theres a downside and its apple shaped though

The only gripe i have atm with pcs is the two pcie slot issue ...........- if you want two high end graphics cards (for rendeirng and video production) and a sfp+ (or faster) card in, you now need to have it janky in your case or go eypc.

S&P credit rating downgraded to BBB....

https://www.reuters.com/technology/sp-global-downgra...

In reply to a post by Pheasant:

I fully expect them to break up, piecemeal like going forward. A sad slow decline.

Yep, very sad. The removal of the only person with good ideas for restructuring will be the day it ended, and the board missed it. The shareholders will remove the board in 3 years when its obviously the new Titanic, and be far too late.

Where the bulk of on-premise compute CPUs comes from will be interesting, AMD doesn't have the scale, and even cloud vendors have been buying lots of Intel x86/64... so maybe AMD will scale production to keep the x86/64 market (mostly cloud or on-premises servers not end user compute) going for 5 years whilst everyone migrates.

Depends if the Qualcomm / ARM lawsuit goes anywhere too!

In reply to a post by jchamier:

Depends if the Qualcomm / ARM lawsuit goes anywhere too!

Steps of the courthouse I expect, where they'll agree to settle.

In reply to a post by Pheasant:

Steps of the courthouse I expect, where they'll agree to settle.

Yes, it has to be only money.

Holy carp. I just read this. Clearly the joint "interim-CEO's" have ablsoutley zero clue or influence here....

https://www.theregister.com/2024/12/13/intel_foundry...

Can kicker:
"As far as, does it ever fully separate? I think that's an open question for another day," Zinsner said in response to a question regarding the role that 18A's success would have on the decision."

Not me Guv:
"Pragmatically, do I think it makes sense that they're completely separated and there's no tie? I don't think so, but someone will decide that," Holthaus said."

Faaaaaaar out. They are utterly clueless.

In reply to a post by Pheasant:

Faaaaaaar out. They are utterly clueless.

Yep, I can see the useless board replacing the useless co-CEOs in less than 6 months. Spiral of doom.

In reply to a post by Pheasant:

Have a read if you have 10 minutes. There’s nothing there that one could really argue with, despite being slightly hyperbolic; it is after all an opinion piece. I certainly wouldn’t argue with the conclusions, even if some of the details aren’t perfect.

Intel have been running on “milk x86 autopilot” for way too long (decades!!) and it’s all finally caught with them.

I fully expect them to break up, piecemeal like going forward. A sad slow decline.

finally read the article, mostly stuff i knew of, but it missed out so many of the server misteps and the 14+++++++++++++++++++++++ process mess

The thing thought and what will kill intel is it has NO noticable ai products from a casual person's point of view. And the Old "you can't go wrong with intel" or "you will never be fired for buying intel" is no longer true.

In reply to a post by Taras:

The thing thought and what will kill intel is it has NO noticable ai products from a casual person's point of view. And the Old "you can't go wrong with intel" or "you will never be fired for buying intel" is no longer true.

This may be the case from the stock market and shareholder point of view, but business in UK (and Europe) is apparently being very very wary of AI. Talk to some CIO's / IT Directors and many of them are keeping well away from general usage AI tools for their staff until there is some test cases on liability. Look at the silly mess Apple's tool caused with the BBC notification alert this week.

More vertically integrated AI tools are more likely to be purchased, such as systems to assist helpdesk agents, or those working on insurance claims.

The future is not remotely clear, if money will be made in AI.

AI software wise, GPDR etc, is a minefield but that should not stop intel from developing devices (yes i know Intel has names stuff as AI) from a chip maker point of view, they are lagging

In reply to a post by Taras:

AI software wise, GPDR etc, is a minefield but that should not stop intel from developing devices (yes i know Intel has names stuff as AI) from a chip maker point of view, they are lagging

I saw their latest mobile CPUs have similar NPU / TOPS capabilities to the Qualcomm Elite and Apple M4 arm based CPUs.

In reply to a post by jchamier:

In reply to a post by Taras:

AI software wise, GPDR etc, is a minefield but that should not stop intel from developing devices (yes i know Intel has names stuff as AI) from a chip maker point of view, they are lagging

I saw their latest mobile CPUs have similar NPU / TOPS capabilities to the Qualcomm Elite and Apple M4 arm based CPUs.

just read this - and i think its no means a perfect read https://www.windowscentral.com/hardware/laptops/what... the premise that intel are scrabbling for ai marketing is that they describe their total unit tops count..

Tops in itself doesn't matter, but what instruction sets are being done. Hence i think AI compute benchmarks are a good year or two.


note: i had to completely redo my reply because that article and my comments were junk because of it .

In reply to a post by Taras:

And the Old "you can't go wrong with intel" or "you will never be fired for buying intel" is no longer true.

The original saying was I believe, “No one ever got fired for buying IBM”

Mmmmm. Funny old world 😂

In reply to a post by Taras:

just read this - and i think its no means a perfect read https://www.windowscentral.com/hardware/laptops/what... the premise that intel are scrabbling for ai marketing is that they describe their total unit tops count..

Its a good article, and shows this measurement is as flawed as all the others in the IT industry; but it does explain that MS use the measurement for Copilot+ certification. Whatever that actually means for the end user.

More woe for Intel with the release of Gartner's semiconductor market update:

https://www.theregister.com/2025/02/04/intel_chip_re...

https://www.gartner.com/en/newsroom/press-releases/2...

Intel in talks to sell chip manufacturing to TSMC

Followed by the news yesterday / today…

Broadcom reportedly investigates acquiring Intel’s chip design biz

You can see why selling fabrication is an interesting idea, become a design house... as Intel are already using TSMC to manufacture recent laptop CPUs.

Will be end of an era though!

Meanwhile a famous design-only CPU house, none other than ARM, announced last week that they will be be building and selling their own chips (probably fabbed by someone like TSMC) for the first time.

Sounds like they want to target the AI and data centre market.

In reply to a post by Pheasant:

Sounds like they want to target the AI and data centre market.

Agreed, and their ownership changes mean they likely want to increase revenue. Perhaps RISC-V will become common

Masayoshi Son wants his ROI. Be interesting to see if RISC-V will make it commercially. It’s a long old road.

What is interesting is that Intel is now in the same situation fab wise, that AMD was in 2009 and sold them . For intel i think its more embarrassing to them given that they saw what happened to amd and how owning their own fab issues caused performance issues for many.