back to article Intel's foundry business bled $7B in 2023 with more to come

Revenue at Intel's foundry business declined in 2023, leading to a $7 billion operating loss, and CEO Pat Gelsinger says this year could produce even nastier numbers as he revealed a reorg to help the chipmaker behave more like its rivals. The reorg will see Intel report results from the following segments: Client Computing …

  1. Lurko

    Nothing like a reorganisation

    Seems to me that the problems are deeper rooted than just being a matter of carving foundries into a standalone business unit, and unless Intel Foundry has costs the same or lower than TSMC, UMC etc then the necessary third party volumes won't appear. And if they don't, utilisation falls, Intel fails to hit the plan, and investors start to push Intel to offload the foundry business - just as AMD spun out GlobalFoundries many years back. Given the losses to date, I can't say I've much confidence that Intel can make the transition and get the costs down.

    Maybe Intel need to admit that running a foundry is very different to chip design, and demerge now.

    1. Roland6 Silver badge

      Re: Nothing like a reorganisation

      > and unless Intel Foundry has costs the same or lower than TSMC, UMC etc then the necessary third party volumes won't appear.

      Maybe they have their fingers crossed…

    2. Timto

      Re: Nothing like a reorganisation

      If Intel spun off their fabs, the new company would probably fail or opt out of latest development (like Global Foundries) and that would leave TSMC with a virtual monopoly.

      We all know what happens when there is a monopoly.

  2. zipityzi

    But why was Foundry revenue down -31% in 2023?

    Why did more Intel Foundry have more sales / revenue in 2022 than 2023?

    I understand Intel used “modeling” to edit its older financial data to create the split between Intel Foundry vs Intel Product, so maybe 2022 had been previously underreported.

    But 2022 & 2023 were under Gelsinger’s reign. Surely Intel Foundry revenue should at a minimum hold steady year over year, if this plan is working.

    How did Intel Foundry sell less in 2023? Lower profits, sure. Lower revenue?

    1. Snake Silver badge

      Re: But why was Foundry revenue down -31% in 2023?

      According to Reuters, Intel failed to invest in the lastest EUV fab technology and instead 'invested' in outsourcing more production.

      Greater costs, and therefore greater losses, thus ensued.

      Big surprise. Management makes bad decisions (who knew?!! /s) and then cry the blues when their short-sightedness for only good quarterlies comes back to bite them.

      1. Sandtitz Silver badge

        Re: But why was Foundry revenue down -31% in 2023?

        "According to Reuters, Intel failed to invest in the lastest EUV fab technology and instead 'invested' in outsourcing more production."

        The article says that was only last year but it was a decade+ ago when Brian Krzanich led Intel and wanted in-house developed chip making process instead of buying from ASML.

        Perhaps the engineers at Intel were overconfident of their abilities and when they failed to produce sub-14nm* chips, the competition (TSMC etc) passed Intel and are far in the lead now.

        *or thereabouts...

      2. Justthefacts Silver badge

        Re: But why was Foundry revenue down -31% in 2023?

        Why are you so convinced the causality between outsourcing/cost/revenue is that way round?

        Seems more likely: Intel chip designer is (still) a profit centre, if only because of market positioning. Whereas it turns out the Intel Foundry isn’t particularly cost-competitive and has been subsidised by chip revenues. What exactly is wrong with outsourcing under those circumstances? It’s only wrong if you are mortgaging your future, ie you have a well-based belief that Intel Foundry *will* somehow become a profit centre in the future. Based on what?

        In the worst case, if Intel Foundry slowly fails into the background, and Intel have to outsource all foundry, TSMC might be able to charge Intel a little bit more due to monopoly. Is that going to kill Intel? I don’t see it. Every other fabless company is currently labouring under TSMC pricing, and those companies can make it work. What new key competitive edge will Intel *Foundry* have in 10yrs time? If we remain on silicon, maybe there’s another generation after 2nm, maybe not. But probably GAA, TSVs and backside power delivery are the last major shrink technologies; and they’re already here. After that, this is purely process cost optimisation at the factory level. Lights out operation etc. Something that TSMC are extremely good at. Honestly, this is a Heads-Tiny-Win, Tails-Big-Fail situation.

        Or…..maybe there will be a game-changing tech post-silicon. Then indeed the economics changes. But Intel Foundry isn’t a blue-sky R&D company.

      3. DS999 Silver badge

        Re: But why was Foundry revenue down -31% in 2023?

        Foundry revenue was down because supply chain issues from covid meant in 2021 and mostly in 2022 all foundries worldwide were running at capacity and had pricing power. Those issues were largely fixed by 2023 so Intel would no longer have been running its foundry business at capacity.

        Plus keep in mind Intel's "foundry business" is basically Tower semi who they acquired to have older nodes. They currently have very little foundry business in leading edge nodes. Maybe an FPGA or two, which was basically the extent of their entire foundry business pre-covid.

  3. RichardBarrell

    That can't be right

    This sounds like the accounting is just a bit incorrect. Intel made a profit last year on GAAP. If their fabs weren't there then Intel wouldn't have been able to make and sell the ~70% of products whose fabrication isn't outsourced. The rest of the business would not respond well if that went away.

    Maybe they should be outsourcing more but the external fab capacity does not actually exist for them to buy. I bet they are not pricing in the fact that their internal foundry has some advantages like the fact that it can give Intel first priority at all times.

    1. Jon 37

      Re: That can't be right

      If they had planned years in advance, Intel could have paid TSMC to fab their stuff at TSMC's standard rates. Knowing that demand was coming, TSMC could have expanded and built more fabs.

      And Intel did plan years in advance, but for Intel Fab to do the manufacturing. So the fair price for Intel Fab to charge Intel Products is TSMC's standard rates. Because that's what the competition would have charged.

      Put another way: Sure, at the time of supply they had no other supplier available who could cope with Intel's volume. But that doesn't mean the correct price is infinitely high. If the supplier was external, they would have signed a long term supply agreement many years ago, which would have locked in a reasonable price.

  4. JamesTGrant

    Cross charging and a re-org - I guess if you’ve got nothing else. Next year it’ll be no pay raises for the normal employees and contractual bonuses will not get paid out, with a message of ‘we just missed out on some sure-fire deals’ and the the following year, oops, next year Rodney. Then there’ll be a new CFO.

    It’s hard to see Intel culture regaining a technical focus. Sad but inevitable.

    1. Anonymous Coward
      Anonymous Coward

      Sound like they have caught the same nickel and diming culture that Boeing have.

  5. Kev99 Silver badge

    Is that a real, dollars and cents, cash money, loss? Or is it the smoke and mirrors AICPA accounting rules, tax evasion loss? Most likely the latter.

  6. CowHorseFrog Silver badge

    Who is the real thief the intel CEO or the iowa sysadmin, and ask yourself which one should have gotten jail time.

  7. NeilPost

    Missed the mobile market

    “But after a rotten 15 years in which it mostly missed the mobile market, ”

    Intel didn’t not miss the mobile market … it sold its lead, for a pittance to Marvell..

    StrongARM/Xscale flogged to pursue the ridiculous mobile Atom that were over priced , under performant and power hungry.

    Schadenfreude writ large.

    1. druck Silver badge

      Re: Missed the mobile market

      While I agree with you they made a massive mistake selling off their ARM business, there is no guarantee it would have given them much of a foothold in mobile - if you consider technical merit rather than marketing muscle.

      The DEC engineers did a fantastic job to take the low power 40MHz ARM7, add Alpha cache technology and create the 230MHz StrongARM, but left in droves when Intel took over. Intel's development was the XScale, they upped the clock speed to 400 to 600 MHz but it was the worst ARM core ever produced, a long pipeline full of instruction barriers and extra cycles all over the place, pathetic memory bandwidth, and an errata that was longer than the ARM7s entire reference manual - basically the Pentium 4 of the ARM world. ARM very quickly regained the lead in performance with the ARM11, and has never produced anything so inelegant or broken.

      But where did those DEC engineers go you ask? The answer is PA-Semi which was subsequently bought by Apple, and their magic is now behind the extraordinary performance of the M series Apple Silicon.

  8. Binraider Silver badge

    Considering the fantastic profits and order backlogs that other fabs are running, what's specific to Intel that's causing problems?

    Malaysia isn't a bad location for a fab by rights, reasonably cheap power and some skilled labour availability.

    This leads me to think that it might be a case of somewhat cooking the books in order to get tax write offs against other elements of the business.

  9. Michael Strorm Silver badge

    $7 billion here, $7 billion there...

    Pretty soon it adds up to real money.

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