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

There has already been much speculation about who will succeed Pat Gelsinger as Intel chief, but perhaps a more pertinent question is whether the company wants to continue to be a manufacturer or join the ranks of fabless semiconductor firms. The chipmaking giant embarked on a turnaround plan envisioned by the outgoing …

  1. Pascal Monett Silver badge

    Well that explains it

    "TSMC has all this expertise in place, but it took them 20 years to get there "

    Those 20 years are the years the USA spent having stuff built elsewhere and not retaining expertise at home.

    You've had your fat margins, now you're going to have to learn to work again.

    1. Yet Another Anonymous coward Silver badge

      Re: Well that explains it

      But during those 20 years American companies like Apple and Nvidia have been way more profitable than TSMC (or at least had better share price rises) so not worrying about how to actually make things is obviously the winning strategy

      1. Ian Bush

        Re: Well that explains it

        Assuming there is somebody there to make the things

        1. seven of five Silver badge

          Re: Well that explains it

          That is what they keep them for, isn't it? Bloody peasants...

        2. Yet Another Anonymous coward Silver badge

          Re: Well that explains it

          >Assuming there is somebody there to make the things

          No need, we lease the factories back from the Private Equity Fund we sold them to (*)

          We can assume that they are running them more efficiently, while investing in new technology and training and retaining skilled workers.

          * That way it comes under the monthly operating budget instead of the capital account... ping .....

      2. JohnSheeran

        Re: Well that explains it

        It's the typical American problem though. It's a short term winning strategy. The failure of the strategy is that it is self-consuming. It is based on the idea that American consumers will continue to be a huge source of revenue because it's where the bulk of the money is. In the long term it will fail because the continued erosion of American wealth will come to a head when only the upper echelon can afford the products that these companies produce. Nvidia's future may not rely on this sort of thing but Apple's certainly does.

        I'm sure this opinion will get downvoted because it's not a snarky one-liner.

        1. Yet Another Anonymous coward Silver badge

          Re: Well that explains it

          > long term it will fail because the continued erosion of American wealth

          Click here and split your loan shark payments into 4 smaller monthly beatings

  2. Gene Cash Silver badge
    FAIL

    Actual work is too difficult and expensive

    Let's spin that off to someone we can pay to do the actual work.

    Wot? We're bankrupt and out of business? How did that happen?

    1. Yet Another Anonymous coward Silver badge

      Re: Actual work is too difficult and expensive

      We've hit an iceberg sir and are sinking !

      No need to worry, we have spun off the ship to Titanic Buoyancy Services, a completely independent company which is solely responsible for keeping the boat above water.

      We don't need to have any concerns about flotation, we are only responsible for selling tickets

      1. Brian 3

        Re: Actual work is too difficult and expensive

        Aren't they the ones who were clever enough to lease excess bouyancy from the titanic?

        1. Yet Another Anonymous coward Silver badge

          Re: Actual work is too difficult and expensive

          The buoyancy was tokenised and sold as NFTs

          1. Anonymous Coward
            Anonymous Coward

            Re: Actual work is too difficult and expensive

            "I'm sorry, did I hear you say you wanted lifeboats and buoyancy?"

      2. UnknownUnknown

        Re: Actual work is too difficult and expensive

        Contracted Service, M-F 8-6, Eastern Time.

        Provided out of Manila.

        1. CA Dave

          Re: Actual work is too difficult and expensive

          More like provided out of accursed India nowadays, and you can't understand them half the time, as the people hired are the ones with the worst grasp of English combined with the thickest accent possible to where you hang up out of frustration.

          It's funny that Trump wants manufacturing done here in USA when it's the Republican fat cats who sent *everything* overseas anyway 20 years ago to abuse cheap foreign labor paid with cheap foreign currency, for fatter profits in the billions back home.

          1. isdnip

            Re: Actual work is too difficult and expensive

            That's not funny, it's part of the big lie. Trump does more projecting than a movie theater chain. Tariffs won't bring industry back to the USA; it'll just break supply chains for what remains.

            But yes a fair number of my phone calls do end up in the Philippines nowadays, not just India. Not that it makes much difference; they still usually just read scripts.

  3. Anonymous Coward
    Anonymous Coward

    GlobalFoundries worked

    Wikipedia says they're the third biggest foundry. Further Wikipedia research tells me they've not gone beyond 12nm. They may not be making the latest and greatest CPUs, but the majority of chips the world needs aren't cutting edge CPUs.

    1. Yet Another Anonymous coward Silver badge

      Re: GlobalFoundries worked

      So Intel's future is making Arduino clones and 555s ?

      1. StrangerHereMyself Silver badge

        Re: GlobalFoundries worked

        More like ARMv7 MCU's for the car industry. Trouble is that the U.S. has few companies left dabbling in consumer electronics.

        1. Yet Another Anonymous coward Silver badge

          Re: GlobalFoundries worked

          A lot of the imported semiconductors aren't CPUs but just regular power components. I can't see the new US embracing IGBTs

  4. Sceptic Tank Silver badge
    Pirate

    LD A, annual_profit / XOR A,A

    The story of Zilog should have left Intel paralyzed with fear a long time ago.Intel is way, way, way bigger but their story is the same: Zilog also hand only one chip design to sell, and some permutations. But once the market moved on they had nothing left to sell, were bought out, and got disappearing railroad blues.

    1. CapeCarl

      Re: LD A, annual_profit / XOR A,A

      My first Unix system was a Z8000 based box in 1980...The Z8000 was a nice beast to develop code on (even in assembler)...Too bad it faded into obscurity.

      (I recall some caution that using the perhaps 700 cycle biggest divide instruction might cause the processor to miss an interrupt.)

    2. kirk_augustin@yahoo.com

      Re: LD A, annual_profit / XOR A,A

      Zilog had no superior chip fab skills or techniques. Intel does. Intel can make chips for video cards, networking, cellphones, etc., and not be involved with competing CPUs.

      1. Richard 12 Silver badge

        Re: LD A, annual_profit / XOR A,A

        In theory, yes.

        In practice, they haven't. Their chips have been almost exclusively CPUs and support chips for those CPUs.

        They've tried other things, but so far they've failed to get market traction - probably because those products have been starved of resources.

      2. Roland6 Silver badge

        Re: LD A, annual_profit / XOR A,A

        I suspect the reality was slightly different:

        "IBM choose Intel because they wanted a CP/M machine (at least at first) but didn't want a Z80 because Zilog was owned by the Exxon oil company and they felt they couldn't bully Exxon."

        Back then Intel was just another chip manufacturer, until the IBM PC...

  5. Anonymous Coward
    Anonymous Coward

    Enter the VC's

    To asset strip the company, add a pile of debt and exit stage left (rapidly).

    With Trump about to put 25% (or more) tariffs on everything from almost everywhere, the US economy is well and truly heading for a monster crash.

    No one will want to invest when so much of the kit that is needed for a foundry has to come from outside the USA (if you want the best kit)

    1. Yet Another Anonymous coward Silver badge

      Re: Enter the VC's

      >when so much of the kit that is needed for a foundry has to come from outside the USA (if you want the best kit)

      Which will drive sales of foreign chip-making kit, causing a trade deficit and a whole bunch of USD in the hands of ASML/TSMC/etc who will have to find somewhere to invest it. And so drive up the price of US stocks like Intel.

      But that's what you get for electing an economics wonk as president.

      1. Roland6 Silver badge

        Re: Enter the VC's

        No reason for ASML/TSMC/etc to accept USD, with the US potentially tailspinning, there are other currencies available…

      2. CA Dave

        Re: Enter the VC's

        Gotta love it when ignorant people think TSMC, et al, will speculate investment in the US stock market when China has proven all you need to do is buy American debt and make loads more off interest payments with the eventual ability to start making demands.

      3. isdnip

        Re: Enter the VC's

        Economics wonk? Trump's tariff plan has zero support from anybody who knows anything about economics. It also comes down to two words: Smoot-Hawley.

    2. UnknownUnknown

      Re: Enter the VC's

      … and a disdain for the Chip’s Act and Inflation Recovery Act and a stated desire to revoke it, revoke funding and recover funding issued (both likely unlawful/unconstitutional) but no spade stuck in the ground yet.

      SCOTUS: Prez - absolute Immunity for Official Acts.

      D’oh.

  6. Mentat74
    Thumb Down

    Whoever becomes the next CEO of Intel...

    It will be most likely a shareholder-patsy that's going to tank the company in less than 5 years...

    1. Richard 12 Silver badge

      Re: Whoever becomes the next CEO of Intel...

      They're big. Circling the drain could take a decade.

      1. CA Dave

        Re: Whoever becomes the next CEO of Intel...

        Rats still flee a sinking ship even if the ship is massive. Especially after 15% of the rats have been culled already.

  7. JoeCool Silver badge

    Doesn't Gelsinger's firing mean the board has already decided on direction ?

    I'm assuming Pat didn't depart over implementation details of an agreable-to-all plan.

  8. kirk_augustin@yahoo.com

    Spin off the fab

    The people who would be alternative fab customers would not be competitors but instead would be GPU, cellphone cpu, and others that have totally different requirements.

    So spinning off the fab is a great idea. Lithography is the one thing Intel is really good at. CPU design, not so much. In fact, if Intel had kept up on their CPU design, GPUs would not ever have existed.

    They would be redundant.

  9. RagnarokStone

    "The government doesn't want to sink public funds into assets that might then end up in foreign ownership, after all." This should have prevented public funds going to TSMC!

  10. Torben Mogensen

    The US needs a local foundry

    I can't imagine the US military wanting chips that are manufactured in Asia, so for national security reasons, the US will want the Intel foundry to survive on US hands. The CHIPS act mentioned in the article is a step towards this, but it may not be sufficient to ensure survival. So the state may force military contractors to use US-based foundries, with the Intel one as a priority (as it has be most advanced technology). It will not be easy, but I can't imagine USA allowing the Intel foundry to die, even if it means pouring massive amounts of cash into it.

    1. UnknownUnknown

      Re: The US needs a local foundry

      https://breakingdefense.com/2024/12/trumps-threat-to-abandon-chips-act-horrific-idea-raimundo-says/

      1. Roland6 Silver badge

        Re: The US needs a local foundry

        However, for those outside the US it might bring benefits, like forcing the US to engage differently with the world outside of itself…

    2. JoeCool Silver badge

      Re: The US needs a local foundry

      Oh yes, the US definitely needs to add a chip fab to the M-I complex. That will lead to incredible market competitiveness and bleeding edge technology available to everyone.

      1. Ace2 Silver badge

        Re: The US needs a local foundry

        You can argue that it’s a bad idea, but you can’t argue that it isn’t overwhelmingly likely to happen.

  11. StrangerHereMyself Silver badge

    Sinking feeling

    I believe part of the reason why Intel is bleeding cash is that it's selling TSMC produced parts on which it makes little profit.

    BTW I keep wondering what all those 100.000 Intel employees are doing on a daily basis? IMHO there's still a world of opportunity shrinking a workforce that size.

    1. BOFH in Training

      Re: Sinking feeling

      Nvidia is also selling TSMC produced products and is making huge profits.

      AMD is also selling TSMC produced products and is making nice profits.

      Only difference : Different types of chips for different purposes from whatever intel is selling (Nvidia GPUs vs Intel CPUs) or better designs (AMD CPUs vs Intel CPUs).

      I know there is some overlap with both Nvidia / AMD and Intel doing GPUs as well, but Intel is such a small player that I don't think I should count it.

      1. StrangerHereMyself Silver badge

        Re: Sinking feeling

        In the case of nVidia it's obvious: they can command huge prices and markup for their products.

        AMD is much smaller and has less overhead and therefore can make a profit. But it's being held back by TSMC's limited production volume. It can therefore never grow beyond 25% market share and that will keep Intel in the lead indefinitely.

  12. Anonymous Coward
    Anonymous Coward

    One does not simply...

    ...build a chip fab.

    The major problem is the supply chain for all the materials. Currently, China dominates there and it would be very difficult for the US to strong arm the supply chains into changing. The US is at a massive geographic disadvantage when it comes to supply chains it is essentially in it's own hemisphere. Even if the US can secure the raw materials, anything manufactured there will always been orders of magnitude more expensive than in China / India / Philippines / Taiwan / Thailand etc.

    There is also the other elephant in the room...labour costs.

    If the US magically manages to fire up some fabs and the price is somehow competitive, then that tells us a couple of things. 1) We've been ripped off for years by the likes of Intel, NVIDIA, AMD etc. 2) It's not as difficult to produce chips as we've been led to believe.

    I suspect by the time the US has some fabs fire up, China will have caught up in terms of their own chip design and we'll have some kind of price war.

    Or mass produced quantum chips are closer than we think and those will be made exclusively in the US for a while which will render high end Chinese product obsolete.

    1. Yet Another Anonymous coward Silver badge

      Re: One does not simply...

      The USA has (had?) a monopoly on the quartz used to make silicon furnaces, In fact it all came from one mine

      It can probably find enough Si, P and B in the continental USA

    2. Anonymous Coward
      Anonymous Coward

      Re: One does not simply...

      Yeah no worries about a sudden US uprising there.

      I was sending resumes around and some of the miserable wages they're offering for technical, credentialed work is actually disheartening. Interviewed at the closest one: no engineer roles but I could start as a manufacturing tech at $16/hour to work on the line (after obtaining a clearance) while Burger King offers $14 to start. This ignores the shouting match the engineer got into with another manager over 'algae into the wash tanks' while I was getting the tour.

      Another one had a manager quit mid-my interview. Another shop wanted to put me on salary, but on the floor, as a "manufacturing assistant."

    3. Tron Silver badge

      Re: One does not simply...

      99% of the chips *needed* are not 'high end'. We've been sold excess computing power for decades now. The PC you are typing your comments to El Reg on could run a manned mission to Mars. You could be using a first gen Pentium PC and it would work just as well. The AI bubble will eventually deflate. It's done better than the Metaverse, but all trends pass.

  13. JRStern Bronze badge

    capex

    Gelsinger lost track of the total capex budgets.

    Gelsinger put too many things on the table at once.

    Gelsinger couldn't fix Intel culture overnight, maybe the biggest challenge of all.

    At this point it comes down to the 18a process, if they can get that to production yield by July 2025 Intel survives.

    Else, I dunno.

    I think they want to keep the domestic fabs. They may need to get more gubmint money to make it work. More this year, more every year.

    AND let's see how the US fabs from TSMC and Samsung work out - they may have many of the same problems working to yield and cost in the USofA.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like