Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
Intel

Slashdot Asks: What Happened To Intel? 120

Intel's board of directors ousted CEO Pat Gelsinger after losing confidence in his ambitious turnaround strategy. The move comes as Intel posted significant losses, including $16.6 billion in Q3 2024, its worst quarterly result ever. Under Gelsinger's leadership, Intel struggled to compete in the AI chip market dominated by Nvidia, while facing manufacturing challenges and declining data center revenue.

Analysts suggest the board may be considering splitting off Intel's foundry business, though such a move could face scrutiny from the U.S. Commerce Department due to $8 billion in CHIPS Act funding. The Verge adds: But Moorhead and Creative Strategies analyst Ben Bajarin both believe Gelsinger's departure was so sudden, it can't simply have been the straw that broke the camel's back. "There must have been a decision the board made that he was not going to stick around for," Moorhead tells me.

His hunch: Intel's board may want to split off its foundry business entirely, above and beyond the spinoff that Gelsinger already announced, turning Intel into a company that simply designs chips like its direct rivals.
This discussion has been archived. No new comments can be posted.

Slashdot Asks: What Happened To Intel?

Comments Filter:
  • by Growlley ( 6732614 ) on Tuesday December 03, 2024 @08:34PM (#64988979)
    next,
    • by 2TecTom ( 311314 )

      classism breeds corruption which produces incompetency

    • by Roger W Moore ( 538166 ) on Tuesday December 03, 2024 @11:11PM (#64989211) Journal
      I think it is more the general life cycle of any human institution like a company. Once you get so large and far ahead of the competition you either get complacent and/or too big to change and adapt allowing the competition to catch up and overtake or you get drunk on success and start doing crazy stuff.
      • by juancn ( 596002 )

        In a sense, you're right, but there are many counter examples, such as Nintendo which since 1889 it has reinvented itself many times.

        It is possible to get out of such a rut, Apple did it at least once too.

        It's a leadership problem. Someone has to push the change and see it through.

        • In a sense, you're right, but there are many counter examples, such as Nintendo

          I would argue that those are not so much counter examples so much as evidence that some institutes are more resilient against these two dangers and so last longer.

      • by kackle ( 910159 )
        I agree. I'd add/wonder about the required resistance to external (financial) vultures who look to infiltrate said company, but their goal is dismantling it instead for personal profit.
    • by gweihir ( 88907 )

      That nicely sums it up. Also refer to Boeing or (still in an earlier stage of this), Google or Microsoft.

    • next,

      And all of the other companies that are prospering using these practices are prospering because... why?

      • Wouldn't that imply that it simply doesn't matter? This list of companies went "woke" and are losing money, this other set of companies did a similar thing and are doing well. My natural cynicism says that hiring practices are ridiculous no matter what ideology drives them and good employees are hired more by accident than deft selection. Perhaps then the problem with America is that they allowed HR to be gatekeepers of staffing and that in the modern world they have a DEI bias is largely irrelevant when it
      • because they are in different stages of the cycle , but sooner or later short termism from quarter to quarter which is the dogma currently will get them. It allows no room for a mistake and is basically self cannibalism .
    • Comment removed based on user account deletion
  • by BrendaEM ( 871664 ) on Tuesday December 03, 2024 @08:35PM (#64988981) Homepage
    During the Bulldozer period, HPDT/Workstation/Gamers had to buy whatever marginally improved chip intel swept from the edge of the desk--for gouged prices. When AMD got their game back, and then Intel was 2nd, no one shed a tear.
    • by jhoegl ( 638955 ) on Tuesday December 03, 2024 @08:58PM (#64989017)
      AMD destroyed intel multiple times over the past 20 years, starting with 64 bit processing. Intel has held dominance only because of its desktop deals.

      Intels grip on them, probably with sweetheart deals to increase desktop profit of large prebuilts, had been propping them up in that market. Which maybe lead to their shitty 13/14 gen CPUs being shitty failure rate chips and them trying to hide it for up to 2 years.

      This on top of the CEO that talked about how sweet their deal was with a chip manufacturer, causing the chip manufacturer to basically stop the deal and charge full price to Intel.

      I would say that gluttony and hubris caused Intels fall in this area, as they basically didnt pay attention to it, didnt care, and thought they had enough dominance to openly gloat and thumb their noses.

      Or in other words, INtel told desktop users "let them eat cake"
      • by fuzzyfuzzyfungus ( 1223518 ) on Tuesday December 03, 2024 @11:06PM (#64989203) Journal
        It seems at best incomplete to say that Intel 'held dominance only because of its desktop deals' given the breakdown in income (and often even more dramatically in margins) between their desktop and their datacenter lines; as well as the fact that they went from making substantial amounts of money to posting losses despite seeing the smallest marketshare reductions in desktop and mobile(both presumably because they do play a lot of hardball with the PC OEMs to encourage them to exclusivity or relegating AMD to the trash tier product lines; but also because the performance and efficiency differences are less dramatic in smaller, lower end parts: AMD's desktop and mobile parts have become vastly more credible with Zen and subsequent iterations; but Intel can knock out single die monolith parts pretty cheaply which makes them much more viable until you get to performance levels where the multi-chip ryzen parts are able to offset their packaging costs better than Intel is able to pump more power into 13 and 14900k dies.

        The shift in desktop(from slightly over 10% AMD in first half of 2017 to slightly over 20% in second half of 2024) and mobile(from 10% to 20% between 2018 and 2024) is certainly noticeable; but it's server CPUs where the shift is most dramatic(~1%in 2017-2018; 24% in 2024) and where the shift in margins is downright brutal: Q2 2024 results for Intel were $3 billion for 76% of datacenter CPUs; AMD saw $2.8 billion on 24%.

        That's an absolute bloodbath on the datacenter side(probably made worse by the number of AI hypebeasts who are spending less on CPUs in order to hit GPUs and 'AI' networking requirements harder): AMD went from basically not existing to making almost half the money between the 7001s and the 9005s.
        • by KlomDark ( 6370 )
          Question: What is 7001s and the 9005s?
          • Sorry, those are model numbers that correspond to 'series' of Epyc server CPUs.

            The 7001 series was "Epyc Rome", based on the zen microarchitecture; started releasing mid-2017. It was followed by the 7002 series (Rome) based on Zen2, in mid 2019, then the 7003 (Milan) Zen3 parts. 4th and 5th gen got a little more complicated; with 4th gen including Genoa, Bergamo, and Siena; with both Zen4 and Zen4c higher core count parts and Sienna's smaller socket and reduced memory controller for smaller systems. 5th
            • by KlomDark ( 6370 )
              Cool, I need to pay more attention to the AMD side of things. I recognize Intel's weird chip identifiers, but not AMD. Which is sad cause I own AMD stock and do not own any Intel stock.
          • by Sique ( 173459 )
            AMD 7001 are the 1st generation EPYC processors (codename Naples) and AMD 9005 are 5th generation EPYC processors (Turin).
      • by tlhIngan ( 30335 ) <slashdot AT worf DOT net> on Wednesday December 04, 2024 @01:20AM (#64989359)

        Intel had dominance because AMD was flailing. They had all the technology, but they were slower and such. Only Sony and Microsoft kept them from going under during the era with the PS4 and Xbox One. And likely Intel as well probably bought a bunch of AMD CPUs to throw away to keep AMD from going under.

        That changed with Ryzen and suddenly AMD was the hotness again with fast chips and good pricing. AMD has had the lead several times - from the K5 era to Athlon series dominating both in price and performance to that weird time with the Pentium 4 Rambus. Then Intel started to dominate again and such.

        It's the cycle of life, and Intel may be fumbling about, but they fumbled about in the past as well. Eventually they'll figure something - likely realizing they need to work on pricing to match the (lack of) performance.

        Competition is good - AMD needs Intel to come around or it'll eventually sit on their ass and release a dud of a Ryzen that's too expensive and too poor performing.

        • by AmiMoJo ( 196126 )

          Intel has been shit for so long that they are having trouble figuring out how not to be.

          Their chips have been mini furnaces since the Pentium 4 era. They have been having hardware defects since the 90s, with the most recent being Spectre/Meltdown and the 13th and 14th gen chips self destructing. They also like to change socket and make everything incompatible every couple of years, so there is rarely any kind of upgrade path like you get with AMD.

          Their manufacturing process is weak too, and AMD is taking fu

        • Intel had dominance initially cos they were better then AMD.

          AMD was better in terms of technology / cost / performance during the Athlon days. Thats when Intel almost killed AMD by basically bribing / threatening OEMs to use Intel exclusively.

          They probably can't do that again since presumably EU and others are watching them (after all Intel got fined a billion bucks if I recall correctly for that bs) and unfortunately they have been stagnant for so long (and having CFO promoted to CEO and intead of doing R

      • by gweihir ( 88907 )

        Essentially, yes. Also note that Intel is essentially a convicted criminal because of massive illegal anti-competitive practices. These a) do not work forever and b) rot your ability to actually compete on merit. Also see Boeing and yet to fall (but it will happen for pretty much the same reasons) Google and Microsoft.

      • Intel has held dominance only because of its desktop deals.

        That is insanely shortsighted. Intel held dominance because it had products people wanted at the price people wanted. At a time when AMD was shipping no iGPU, Intel offered businesses what AMD couldn't. At the time AMD was promising more cores, Intel offered gamers what AMD couldn't (better IPC). Just after the move to 64bit Intel also trumped AMD on efficiency per clock (after behing behind in the P4 days).

        It wasn't until recently (and I mean post pandemic recently) that AMD really had a product offering t

      • AMD destroyed intel multiple times over the past 20 years, starting with 64 bit processing.

        on the desktop side, it was only twice in 2 decades: athlon 64, and the current 3d cache chips (specifically for gaming).

        intel's conroe/core 2 duo annihilated amd and sent them into therapy.
        bulldozer was a joke and eventually amd came back with the zen architecture... but it was only competitive, not top-of-the-line.

        not as familiar with server/enterprise CPUs, but everything ive seen points to epyc/threadripper being well regarded, but not destroying intel's offerings.

    • by UnknownSoldier ( 67820 ) on Wednesday December 04, 2024 @01:47AM (#64989381)

      > had to buy whatever marginally improved chip intel swept from the edge of the desk--for gouged prices.

      Worse, post Bulldozer in the HEDT era the perfect example, IMHO, of Intel price gouging customers was the 50% price drop of the i9 10980XE compared to the previous i9 9980XE! How do you price drop a high-end CPU by 50% over 1 generation if the previous version wasn't price gouging???

      Intel literally held quad-core computers back almost a decade before AMD made "cores for cheap" with a fantastic bang/buck especially with Ryzen and the AM4 platform.

      If Intel isn't careful AMD might end up buying them.

  • Why would anyone buy an Intel CPU over an AMD CPU when team red is so much better?

    • There have been reports of OEMs unhappy with supply, particularly for laptop parts(not really an enormous surprise when AMD has to choose how to divide their chunk of TSMC's attention between client margins and datacenter margins, along with the presumably longer term contracts for the semi-custom APUs in everyone except Nintendo's current-gen consoles.

      That, along with low to midrange desktops, is also where Intel's offerings are at their most credible. The persistent attempts by some vendors to depict a
      • Laptop OEMs are still enjoying deep discounts on 10nm parts from Intel. AMD is charging.a lot for new products like Strix Point (not so much for previous gen Hawk Point).

        Take supply complaints from OEMs with a grain of salt.

        • OEMs are definitely self-interested parties, so their whining could definitely be tactical(though I'm not 100% sure how much movement they'll get out of AMD when there's a fairly sharp difference in battery life on both AMD and Intel sides between Strix Point/Lunar Lake and earlier, so the 10nm parts are a bit harder to present as direct alternatives); but when the question is "why would anyone buy intel?" "Attractive discounts on basically all the 10nm they can sell you" is a pretty good answer if you just
      • There's also a certain amount of "no-one ever got fired for buying Intel", or at least there was. I've worked at a number of places where this wasn't even open to debate, if you wanted a PC you got an Intel PC.
    • Re: (Score:2, Interesting)

      by bemymonkey ( 1244086 )

      Stability on mobile. My Ryzen based Thinkpads have all been shitshows due to a combination of UEFI/EC FW bugs and bad drivers. Things like Windows refusing to reliably wake from standby/hibernate right down to random shutdowns/reboots... this behaviour persisted until enough firmware updates had been released, meaning for the year or two that I used the devices I got about 3-6 months of stability out of each.

      I've never ever had that kind of instability on an Intel based T/X/W/P series Thinkpad.

      • by Kazymyr ( 190114 )

        Interesting. It may be a Lenovo problem, not an AMD problem. Because my HP laptops with Ryzen have been extremely stable and reliable in the past 3 years. Not one single problem.

        • I do agree that it's a Lenovo problem, but I'm assuming there are underlying reasons, such as (speculation) more mature reference designs, better example code and better support from Intel for UEFI FW and EC FW dev. - I find it unlikely that the AMD platform FW dev teams at Lenovo just dropped the ball while the Intel platform teams are simply much more competent...

          My experience has been (and continues to be) ~5 modern Intel Thinkpads with no issues and 3 modern AMD Thinkpads (Ryzen 6xxx gen and up) that al

        • by MikeS2k ( 589190 )

          I've had this with both Intel and AMD based Thinkpad E14's - Gens 2 up until 4 (manage a fleet of about 100) - issues with Windows and standby; problems with the sound, etc - the latest BIOS updates have fixed these problems in both cases. I wonder if Lenovo's QA is a bit sloppy when these newer Thinkpads are released

      • by gweihir ( 88907 )

        Your one data-point is interesting. I raise you one "I have never had this problem on my AMD thinkpads (E14 and a new Idea Pad slim).

    • by jezwel ( 2451108 )
      We've found some software manufacturers refuse to guarantee compatibility with AMD CPUs, and that's because we've found newer AMD designs have something that breaks their software.

      Not a problem unless the software is very niche...

      • That seems shortsighted of them. The errata list for an Intel CPU is often substantial and AMD is no less well endowed in that particular back of the manual information. So it's guaranteed that they are either already implementing or relying on a 3rd party implementing Intel workarounds, so to say "we won't certify AMD because they suck" that just tells me "we are going to ride our one vendor policy into the grave."
    • by gweihir ( 88907 )

      Stupidity, subserviency to the "biggest" player, mental inflexibility. Intel has had the worse offer for half a decade now or longer and across the board.

    • by vbdasc ( 146051 )

      I would buy even a VIA CPU for the right price.

    • I don't know about *so much* better, but marginally more like it.

      Both have good and bad. I prefer AMD right now, and I don't see myself buying Intel for any reason in the next five years. AMD is in the lead, but I in the next five years, could a RISC chip be equally in the lead? Maybe.

      Ryzen seem pretty good in the laptops these days, just as much in the desktops. Why are so many laptops Intel? Don't those sales topple desktops right now?

      Maybe people just need to be more familiar with red stickers on the lap

  • by BrendaEM ( 871664 ) on Tuesday December 03, 2024 @08:57PM (#64989015) Homepage
    While server motherboards can cost $1000, alone, and while they aren't giving away server CPUs, memory is huge expense. AMD has usually let you upgrade the CPU once, saving motherboard and RAM costs, while Intel is basically a one-shot deal. This also affects the choice of parts too.

    If you think that RAM isn't a big deal, then please make arrangements to send me a MB/RAM/CPU combo, to my sig. I need, absolutely need 256GB of RAM, want 512GB or better for doing OpenFoam Computation Fluid Dynamics tests on a home-spun car body. If I had 1-2 TB of RAM, that would get the unbiased mesh down about 1mm, which is at the point of diminishing returns.
    • by zeeky boogy doog ( 8381659 ) on Tuesday December 03, 2024 @09:30PM (#64989073)
      Que? Both Intel and AMD have been on a "multiple generations of CPUs per motherboard socket" routine for a long time now.

      Intel:
      - Sandy Bridge then Ivy Bridge [LGA2011]
      - Haswell then Broadwell [LGA2011-3]
      - Sky Lake then Cascade Lake [LGA3647]
      - Cooper Lake then Ice Lake [LDA4189]
      - Sapphire Rapids then Emerald Rapids [LGA4677]

      AMD:
      - Opteron/Bulldozer [G34 - ZIF1944]
      - Zen, Rome, Milan [SP3 - ZIF4094]
      - Genoa, Bergamo, Turin [SP5 - ZIF6096]

      If you're talking about support for O(1TB) of memory, I assume you're referring to server parts because until pretty recently even most server motherboards didn't implement enough physical memory address lines to support 1TB.
      • Nobody really bought Cooper Lake though.

      • My work has a shared development server from about 2020, with dual xeon cpu's and 1TB of RAM
        It's just an average multi-socket HP server I think.

      • by AmiMoJo ( 196126 )

        Intel typically released a platform, then released a refresh of it the next year, and then it was dead and you needed a new motherboard to upgrade.

        It was artificial obsolescence, and some Chinese manufacturers found ways around it that let them recycle chipsets and CPUs from Intel e-waste.

        • So you want both a processor with 1.5 to 2 times the bus speed and i/o pin count and core count and power consumption, AND for it to be pin compatible with a now two or more generations old motherboard? This has never been a thing from any manufacturer because the idea is ridiculous on its face, as now as it would've been in 1994 if you'd been demanding to be able to plug a Pentium into a 286 motherboard. Get real, and be thankful that the connection standards for peripherals like sata, nvme and pcie are ba
  • Same thing as Google (Score:5, Interesting)

    by phantomfive ( 622387 ) on Tuesday December 03, 2024 @09:00PM (#64989019) Journal
    They had some competent people in charge, a good product. But the leeches started coming in to management, more interested in grabbing what they could than improving the company (or making the world a better place). These people had more skills in backstabbing their coworkers than in running a competent business.

    Once the competent people retired, the leeches took over the place and didn't know how to keep it running. They fired a bunch of their smart technical people, cutting the wheels out from under their bus.

    Really, the amazing thing is that Intel managed to keep going for two decades after that happened. They had some strong inertia.
    • by dziban303 ( 540095 ) <dziban303@NoSpaM.gmail.com> on Tuesday December 03, 2024 @09:22PM (#64989059) Homepage
      Funny how that keeps happening
    • Happened when McDonnell Douglas execs took over Boeing.

      It's mind boggling how someone thought it was a good idea to not fire the management of the failing McDonnell, and keep the Boeing execs.

      • I've seen it happen a bunch of times firsthand. Common story, a startup goes public, and the leeches come, usually looking to join in management. They go around backstabbing, etc, until they ruin the company. It's to the point that I consider the #1 challenge for a startup going public to be managing this problem.

        Of course, the fact that the original management is tired and wants to exit the company often compounds the problem. But the main problem is the leeches coming in. As one small example I give Ada [crunchbase.com]
  • Simple (Score:4, Interesting)

    by Pollux ( 102520 ) <speter AT tedata DOT net DOT eg> on Tuesday December 03, 2024 @09:20PM (#64989053) Journal

    Competition happened.

    Intel's had plenty of blunders over the years. Remember when 2+2 didn't have to equal 4 [wikipedia.org]? Good times, good times.

    But through all their mistakes, there was never anyone who was really poised any threat to overtake them. AMD was the only real competition in the early aughts, but even they discovered how hard it was to really dethrone Intel. Like, impossibly hard.

    But no dynasty lives forever, and eventually their winning streak came to an end. They spent way too long trying to figure out how to shrink their fab below 10nm, and others caught up then surpassed them. Now it's rebuilding time...if they can afford to do so.

    • by necro81 ( 917438 )

      Remember when 2+2 didn't have to equal 4?

      The FDIV bug, thoroughly reported, would not have affected 2+2. It didn't even affect most division calculations. The wikipedia entry you linked to includes this tidbit:

      The severity of the FDIV bug is debated. Though rarely encountered by most users (Byte magazine estimated that 1 in 9 billion floating point divides with random parameters would produce inaccurate results) [ref [byte.com]]

      Doesn't mean they didn't screw up. (As with most things, the way they handled the mi

      • My memory is extremely hazy on this but I remember people saying that it was a rare occurrence and not worth worrying about, but then people started to notice this "rare" thing wasn't that rare in practice. FP calculations aren't composed of wide ranging random numbers, you tend to have a whole lot of calculations within a narrow range of values and if that's the problem range, then you'll get a whole lot of imprecision. And that was ultimately why Intel had to issue the belated recall, when it became demon

  • by locater16 ( 2326718 ) on Tuesday December 03, 2024 @09:34PM (#64989075)
    Still losing money. Still losing in servers to AMD and ARM. Missing the AI boat completely despite spending billions on 2 separate AI startups. And now AMD and Qualcomm are really pushing to muscle in on Windows laptops, the last place Intel really makes money without a lot of competition. Whether silicon fabs work or not the man failed elsewhere given almost 4 years of tenure.
  • by ndykman ( 659315 ) on Tuesday December 03, 2024 @09:37PM (#64989079)

    They are being compared to Nvidia and the massive AI bubble they are riding. They are in the middle of some bad PR and bad execution for sure, but it's really a lot of short pressure from investors that want "all the money now" versus long term growth and investment in fabs and more reasonable investments in AI acceleration. Frankly, when the AI bubble pops, NVidia will be way more exposed to downward market pressure than Intel and AMD.

    They have a very good story on the newest Xeons and has a really good roadmap overall. And AMD is putting good pressure on them to get on top of things.

    Intel would do well to spin the foundry side back in and shore it up and frankly, they should get more government support to do so. Having TSMC be the only source of high end fabrication is not a good thing. Yea, it's protectionist, but in this case, I think that's a good thing.

    • TSMC, Samsung all get huge govt support, for decades. It looks like the only way to compete internationally in the fab race. This has national security implications as well. If TSMCs supply chain dies (due to war etc) then we will want a domestic replacement. Even if theyâ(TM)re only 95 percent as good, they can still take over much of TSMCs domestic business in a few years. The govt is playing the long game. Possibly Akin to ship building. Because if people forget how to build them domestically we a
  • by rsilvergun ( 571051 ) on Tuesday December 03, 2024 @09:40PM (#64989089)
    They couldn't keep up with AMD because they were starved for capital after years and years of taking money and using it for stock buybacks in order to artificially boost share price and line the big shareholders and CEO pockets.

    It also meant they couldn't take any useful risks like outbidding AMD for the PS4 contract. To be fair consoles tend to be low profit margin because as the console life cycle continues there's pressure to make cheaper CPUs and GPUs but still that didn't happen this time around with console prices staying pretty high allowing AMD to consistently make money off the hardware.

    They also dragged their feet getting into the GPU market so they're too far behind to really take advantage of the Bitcoin and AI bubbles. Again they didn't have a lot of money lying around to take risks with because so much money was tied up in stock buybacks.

    There's a good reason stock buybacks used to be illegal and why they should be now. But our economy is going to be such a mess for the next 4 years it hardly matters
  • by thesjaakspoiler ( 4782965 ) on Tuesday December 03, 2024 @10:26PM (#64989147)

    That has been the core problem for the past 25 years or so.
    Intel was always pushing second class GPUs with a lot of buzzwords.
    But when it came down to performance, they never managed to build something that came close to nVidia or AMD.

  • by Matt_Bennett ( 79107 ) on Wednesday December 04, 2024 @08:46AM (#64989895) Homepage Journal

    (Purely technical perspective, don't really know much about the business side.)

    I may have some good insight here- worked for a major computer manufacturer in networking. For at least 20+ years (30?), their strategy has been to pull as much computing back in to the CPU as possible. There are some good reasons- high bandwidth, low latency. (and problems, heat and analog performance). They kinda did a "pshaw" with all the video stuff, and maybe a half-hearted attempt, but when the algorithms broke out into the GPU, they were in trouble.

    Overall, I expected that the concentration on the CPU itself was the downfall. Yes, Intel is active in many other technology areas, but their focus on all these areas is that they are peripherals of the CPU, connected as close as possible to the CPU. When the peripheral does not push its work to the CPU, it wasn't prioritized.

  • Like everything, there are multiple ways to look at the problems. But, there is a fundamental design side, and then there is the fabrication side, and Intel really got lost when it comes to design going back 15 years now. If you think about it, you can take a design, and put it on almost any fab process. There will be a lot of differences when it comes to the fab process, clock speeds, how much power things can handle before things break down, power draw, and more, but at the heart, you have to think

  • No Andy Grove, no Intel. This is repeated innumerous times.

  • by Gavino ( 560149 ) on Wednesday December 04, 2024 @04:47PM (#64991159)
    His name? Elon Musk.

The light of a hundred stars does not equal the light of the moon.

Working...