Former Intel CEO says splitting Intel isn’t good for the U.S.

Intel Logo
(Image credit: Getty Images)

Intel’s financial situation has recently been in trouble, mainly due to the massive losses of its foundry division. Because of this, several reports say that the company is considering spinning off manufacturing as a completely different entity, similar to what AMD did with GlobalFoundries. However, former Intel CEO Craig Barrett wrote on Fortune that doing so would hurt Intel and impair America’s goal of semiconductor leadership.

Barrett argues that chip manufacturers require massive investments to remain competitive. Only three chipmakers — Intel, Samsung, and TSMC — have the revenue to sustain the research and development needed for future products. Intel’s design arm would likely survive this split, but it’s an open question if Intel’s foundry business would last.

He compared this to the AMD-GlobalFoundries split in 2008. Today, AMD is doing well in the CPU field with its Ryzen processors, and TSMC produces many of its chips. On the other hand, GlobalFoundries has fallen behind TSMC because it does not have any differentiating technology that would allow it to be on the leading edge of semiconductor manufacturing.

With its limited resources, GlobalFoundries couldn’t invest more in research and development, lagging behind its competitors. Potential customers then went with foundries that had more advanced technologies, limiting its production and revenue, which, in turn, led to even less cash for R&D. The former Intel CEO says that the same thing could happen to Intel’s manufacturing division if the company (or the U.S. government) splits it off as an independent company.

If that were the case, the U.S. would likely depend on TSMC or Samsung to produce its latest chips. Although Washington has encouraged these companies to set up shop in the mainland U.S. through the CHIPS Act, with the latter planning to start regular production by early 2025, Barrett says these companies are still foreign suppliers. Furthermore, the headquarters of these companies are in countries with external threats and could serve as a flashpoint for a major conflict.

Instead of breaking up Intel, he suggests that Intel and the U.S. government should focus on what matters most for semiconductor companies — technological leadership. If the company focuses on delivering superior performance over its competitors, Intel would be able to turn itself around and once again be the global leader in chip manufacturing. The former CEO even praised the current head of Intel, Pat Gelsinger, for being on the right track for Intel, especially with the recent reveal of the Clearwater Forest Xeon chip that’s built on the 18A process node.

Barrett also said Washington must do its job to stay ahead of the semiconductor race. The U.S. has invested more in the semiconductor industry in the past year than in the last 28 years combined, but he says it should do even more, especially in academic research. While establishing the National Semiconductor Technology Center (NSTC) is already a good start, he also points out that its five-year budget is smaller than Intel’s annual R&D expenditure.

Mr. Craig Barrett says that Intel has been through a similar situation before—when the dot-com bubble burst and caused the entire tech industry to melt down. Despite all that was happening, Intel kept investing in its R&D programs, and when the dust finally settled, the company was in a stronger position than ever before. The former Intel chief believes that it should do the same again this time.

Jowi Morales
Contributing Writer

Jowi Morales is a tech enthusiast with years of experience working in the industry. He’s been writing with several tech publications since 2021, where he’s been interested in tech hardware and consumer electronics.

  • hotaru251

    Instead of breaking up Intel, he suggests that Intel and the U.S. government should focus on what matters most for semiconductor companies — technological leadership.

    tl;dr "Give Intel more $$ to try and brute force their way to dominance"?

    Gov shouldnt have to fund a company as if they cant do it themself why doesnt Gov just fund a new company themself and own it entirely?
    Reply
  • ThisIsMe
    hotaru251 said:
    tl;dr "Give Intel more $$ to try and brute force their way to dominance"?

    Gov shouldnt have to fund a company as if they cant do it themself why doesnt Gov just fund a new company themself and own it entirely?

    That’s one way to interpret it. I mean if the government sent everyone a check for some amount of money, I’m certain you’d cash it.

    However, maybe he was suggesting that the government should not support foreign organizations nor domestic organizations that don’t actually make anything and rely solely on foreign organizations to make their products. Instead, it should support its only leading edge domestic manufacturing organization. You know, like what he literally said.

    About your last thought, it can’t.
    Reply
  • ezst036
    The CHIPS act has become nothing but routine crass corporate welfare.

    Gimme, gimme, gimme. So insulting. Yay for the lobbyists though.
    Reply
  • thestryker
    hotaru251 said:
    tl;dr "Give Intel more $$ to try and brute force their way to dominance"?

    Gov shouldnt have to fund a company as if they cant do it themself why doesnt Gov just fund a new company themself and own it entirely?
    Ah yes the approach that doesn't take into account subsidization that happens globally or the way wall street works. In a rational world you'd be right, but that isn't the world we live in.

    Pat laid out a plan when he became CEO and everyone seemed to be onboard. Intel split out fabrication financials (it's obvious this was a money loser when they're having to buy extremely expensive equipment while maintaining existing and not being able to use the new until it's ready which means idle fabs) and suddenly everyone is freaking out because foundry isn't making money. Intel's valuation tanks and they start doing typical corporate stuff to stem the bleeding even though it's usually bad for long term health. If none of this happens the fab delays probably don't happen either.

    Government subsidies to get fabs built out with any speed is basically necessary at this point. The thing that needs to be done is putting guardrails on those funds to make sure they're being used to do exactly that rather than just line shareholder pockets.

    There is no way spinning off the fabs today ends with a leading edge fab down the road. If 18A ends up being a success and Intel brings in enough customers to fill the fabs they have that's when the conversation about spinning off starts to make sense. It isn't until the "12" node with UMC is completed that it really makes sense though as that will be able to leverage all of the existing DUV equipment for an industry standardized node.
    Reply
  • hotaru251
    thestryker said:
    Ah yes the approach that doesn't take into account subsidization that happens globally or the way wall street works. In a rational world you'd be right, but that isn't the world we live in.

    Pat laid out a plan when he became CEO and everyone seemed to be onboard. Intel split out fabrication financials (it's obvious this was a money loser when they're having to buy extremely expensive equipment while maintaining existing and not being able to use the new until it's ready which means idle fabs) and suddenly everyone is freaking out because foundry isn't making money. Intel's valuation tanks and they start doing typical corporate stuff to stem the bleeding even though it's usually bad for long term health. If none of this happens the fab delays probably don't happen either.

    Government subsidies to get fabs built out with any speed is basically necessary at this point. The thing that needs to be done is putting g guardrails on those funds to make sure they're being used to do exactly that rather than just line shareholder pockets.

    There is no way spinning off the fabs today ends with a leading edge fab down the road. If 18A ends up being a success and Intel brings in enough customers to fill the fabs they have that's when the conversation about spinning off starts to make sense. It isn't until the "12" node with UMC is completed that it really makes sense though as that will be able to leverage all of the existing DUV equipment for an industry standardized node.
    issue is who would use their fabs (and at scale enough to be profitable) when even Intel wont make their own CPU's there and instead outsource them to TSMC?

    By not making your own cpu's you are admitting you can't do that great yet and why would anyone want to trust that?
    Reply
  • Alvar "Miles" Udell
    Unlike what AMD did, spin off their foundries and sell the majority of ownership to the UAE (and are now public), Intel is wanting to spin off the fabs as a wholly owned subsidiary, like Intel Ireland (now 49% owned by Apollo), or vastly majority owned subsidiary, like Mobileye. It'd be better for Intel as they could acquire additional funds by making them public, it'd make them more attractive to contractors since they'd be just fabricators, and it'd make a takeover or merger easier as there would be no assets to shed.
    Reply
  • thestryker
    hotaru251 said:
    issue is who would use their fabs (and at scale enough to be profitable) when even Intel wont make their own CPU's there and instead outsource them to TSMC?

    By not making your own cpu's you are admitting you can't do that great yet and why would anyone want to trust that?
    This is nonsense. LNL is the only product which was designed not using an Intel node. ARL was a decision based on a node cancelation.

    The only manufacturing process Intel isn't using is 20A which they claim is due to focusing on 18A which makes sense (whether true or not). MTL is Intel 4, SRF/GNR are Intel 3 and CWF/PTL are 18A. The reason canceling 20A makes sense is that with MTL Intel has to keep EUV machines making a singular product on Intel 4 which means those can't be used for Intel 3 and the same would have been the case with 20A/18A. Intel is trying to open up leading edge foundry services with 18A and having capacity available is paramount for this.
    Reply
  • ToBeGood
    ezst036 said:
    The CHIPS act has become nothing but routine crass corporate welfare.

    Gimme, gimme, gimme. So insulting. Yay for the lobbyists though.
    According to Intel Financial Report, Intel still yet received 1c from the CHIPS act as of Q2 2024.

    <Mod Edit - delete off topic political>
    Reply
  • LordVile
    Splitting nvidia would be more beneficial at this point
    Reply