A Reuters report paints a rough picture, but Intel 18A still appears to be on track.
An exclusive report from Reuters on Wednesday appeared to paint a picture of turmoil at Intel‘s (INTC -3.33%) nascent foundry. According to multiple sources, test wafers for semiconductor giant Broadcom that had run through Intel’s upcoming Intel 18A manufacturing process came back flawed enough for Broadcom to conclude that Intel 18A was not ready for high-volume production.
Water is wet
The Intel 18A process isn’t ready for high-volume production because… it’s not ready for high-volume production. Intel’s timeline is to have the process, meant to beat foundry leader TSMC technologically, ready for manufacturing by the end of the year and ready for high-volume manufacturing for external customers a bit later in 2025.
It shouldn’t be a surprise that a manufacturing process that’s not ready for mass production is producing subpar wafers. In response to the Reuters piece, an Intel spokesperson said that Intel 18A is healthy, yielding well, and on track for high-volume manufacturing next year. A Broadcom spokesperson added that the company’s evaluation of Intel 18A has not yet been concluded.
Intel also disclosed some information at a recent conference about how the Intel 18A process is evolving, and the numbers look good. Intel CEO Pat Gelsinger said that the defect density is now below 0.4 defects per square centimeter. A defect density below 0.5 is generally considered good, so Intel 18A appears to be in good shape. With high-volume production still months away, Intel has time to drive that metric even lower.
Intel now has a dozen customers using its Intel 18A manufacturing toolkit and plans to launch its own products on Intel 18A next year. Those internal products include Panther Lake PC CPUs and Clearwater Forest server CPUs.
Still a long road ahead
Based on what Intel has said and the numbers it’s provided, the Intel 18A process appears to be moving along according to schedule. Capacity remains an open question. Once Intel 18A is ready for high-volume production, will the company be capable of handling orders from a large customer like Broadcom? Or will it take a considerable amount of time for Intel to scale up production to meet both its own and its customers’ needs?
While the headline of the Reuters article, which came soon after a disastrous second-quarter earnings report from Intel, suggests that the company’s all-important Intel 18A process is in trouble, that doesn’t appear to be the case. Intel 18A is still multiple quarters away from being used by external customers for high-volume production, so the fact that it’s not ready right now isn’t really a story.
Still, Intel must win customers for Intel 18A, and it’s hard to do that when the process isn’t ready yet and the company has no track record as a foundry. Intel scored Microsoft as an Intel 18A customer for an undisclosed chip earlier this year, but other customer announcements have been few and far between.
It’s going to take time for Intel’s foundry business to generate meaningful external revenue as customers become comfortable with Intel as a foundry. If Intel 18A is performing like the company says it is, the foundry business should be in good shape in the long run.
Timothy Green has positions in Intel. The Motley Fool has positions in and recommends Microsoft and Taiwan Semiconductor Manufacturing. The Motley Fool recommends Broadcom and Intel and recommends the following options: long January 2026 $395 calls on Microsoft, short January 2026 $405 calls on Microsoft, and short November 2024 $24 calls on Intel. The Motley Fool has a disclosure policy.