CPU Failures Hurt Intel's Bottom Line

Intel revealed during its Q4 2016 earnings call that unspecified CPU failures have affected its bottom line. The company enjoyed record Q4 revenue and a record $59.4 billion in overall 2016 revenue, but CPU failures reduced the Data Center Group's financial performance. 

Intel's Robert Holmes Swan, the new CFO and executive vice president, vaguely defined the failures as he explained why the Data Center Group's (DCG) profit declined 14% year-over-year to $1.9 billion. Swan listed the 14nm transition and a one-time long-term cross-licensing agreement as issues that contributed to the decline, but also mentioned undefined CPU failures as a drag on financials (via Seeking Alpha):

"But secondly, and a little bit more significant, we were observing a product quality issue in the fourth quarter with slightly higher expected failure rates under certain use and time constraints, and we established a reserve to deal with that. We think we have it relatively well-bounded with a minor design fix that we're working with our clients to resolve. So, those two one-timers in the fourth quarter weighed on DCG [Data Center Group] margins, and we do not expect that to continue in 2017."

Intel's Data Center Group produces numerous products that fall into a wide range of segments, such as CPUs, networking, and storage, so we followed up with Intel for more clarity on the nature of the failures. Intel replied that the failures are limited to the Intel Atom Processor C2000 product family, which is used primarily for storage and networking infrastructure.

Intel has launched a concerted campaign to extend its data center CPU stranglehold (it occupies over 99% of the world's servers) to other lucrative "adjacencies," such as the networking and storage segments where it has relatively low market penetration. Due to Intel's efforts in these areas, it has reduced the DCG's reliance on the traditional enterprise CPU market, which now only accounts for 50% of its earnings. Unfortunately, the CPU failures are confined to two of the areas that Intel defines as key growth drivers.

Intel indicated that it has an immediate workaround that addresses the issue, as well as a "minor" silicon fix that solves the problem entirely. As noted in the earnings call, the problem is of enough import that Intel had to establish a reserve to address the issue, so frankly, Intel may be downplaying the problem. Any new silicon revision requires significant expenditures and engineering effort, as proven by the drag on DCG's bottom line.

Intel doesn't consider the failures to have a long-term impact on its financial performance, and, for now, the still-undefined issue appears to be limited to the C2000 Atom family. We're following up for more specific information on the failures, and will update as needed.

Paul Alcorn
Managing Editor: News and Emerging Tech

Paul Alcorn is the Managing Editor: News and Emerging Tech for Tom's Hardware US. He also writes news and reviews on CPUs, storage, and enterprise hardware.

  • gggplaya
    Giving room for AMD to take some marketshare with their new zen based CPU.
    Reply
  • dgingeri
    Does it say CPUs specifically were the cause? I don't see it specify CPUs in what is quoted.

    I do, however, know of a significant problem with their X710 10Gb CNA that cause significant problems early last year. Many of the X710 CNAs had a problem with stalling and corrupting data under heavy use, which rendered it useless for a product my company was working on at the time. We ended up going back to the X520 in order to get our product out because we couldn't trust the X710. I'm fairly certain that impacted their bottom line pretty well. My company at the time used returned about 300 of those X710s to Intel for them to analyze for the failures, which is no small change. (That was just the test phase, not production, or there would have been many more.)
    Reply
  • 3ogdy
    19234025 said:
    Does it say CPUs specifically were the cause? I don't see it specify CPUs in what is quoted.

    I do, however, know of a significant problem with their X710 10Gb CNA that cause significant problems early last year. Many of the X710 CNAs had a problem with stalling and corrupting data under heavy use, which rendered it useless for a product my company was working on at the time. We ended up going back to the X520 in order to get our product out because we couldn't trust the X710. I'm fairly certain that impacted their bottom line pretty well. My company at the time used returned about 300 of those X710s to Intel for them to analyze for the failures, which is no small change. (That was just the test phase, not production, or there would have been many more.)

    " Intel replied that the failures are limited to the Intel Atom Processor C2000 product family, which is used primarily for storage and networking infrastructure."
    Reply
  • dgingeri
    19234049 said:
    19234025 said:
    Does it say CPUs specifically were the cause? I don't see it specify CPUs in what is quoted.

    I do, however, know of a significant problem with their X710 10Gb CNA that cause significant problems early last year. Many of the X710 CNAs had a problem with stalling and corrupting data under heavy use, which rendered it useless for a product my company was working on at the time. We ended up going back to the X520 in order to get our product out because we couldn't trust the X710. I'm fairly certain that impacted their bottom line pretty well. My company at the time used returned about 300 of those X710s to Intel for them to analyze for the failures, which is no small change. (That was just the test phase, not production, or there would have been many more.)

    " Intel replied that the failures are limited to the Intel Atom Processor C2000 product family, which is used primarily for storage and networking infrastructure."

    Ah, didn't see that. Well, I'm fairly certain the X710 thing didn't help either.
    Reply
  • blazorthon
    That makes sense. The little CPUs are often put in systems that only barely meet minimum requirements. If there's a small error made in what the chips need to avoid heat damage, then they fry.
    Reply
  • damric
    No Skylake CPUs that used JEDEC DDR3 were listed as failing.
    Reply
  • dstarr3
    19234018 said:
    Giving room for AMD to take some marketshare with their new zen based CPU.

    Except AMD have no CPU vying for market share with Intel's Atom processors, like the ones this article states are affected.
    Reply
  • brucewhain
    ...to other lucrative "adjacencies,"
    Reply
  • photonboy
    Blazorthon,
    Nobody said they failed due to heat. In fact, I doubt it considering they have a "workaround" so it could be some errata bug that is fixed with a firmware update.
    Reply
  • slidai
    I've always had Intel chips. When I built my new gaming rig a couple of months ago, after a day of use, it crashed on start up over and over. Turns out it was the cpu.
    Reply