Seeking Alpha

WSJ: IBM mulls sale of semiconductor manufacturing ops

  • IBM (IBM) is considering a sale of its semiconductor manufacturing operations, which produces chips for its high-end servers and for external customers, the WSJ reports.
  • However, IBM would keep its chip-design operations; alternatively, according to an earlier FT article, IBM could keep the whole business and seek a partner with which to create a joint venture. The FT reported about the prospective sale but didn't distinguish between the production and design parts of the division.
  • There was no word on the value of the business, although Bernstein reckons it earned revenue of $1.75B in 2013 and made $130M in pretax losses.
  • The reports come after IBM said last month that it had agreed to sell its low-end server division to Lenovo for $2.3B.
  • Evisioneering analyst Rick Doherty said he'd be "shocked" if IBM did sell the semiconductor operations. "Take away the Silicon part, and IBM may not be the tech giant it is 10 years from now," Doherty said.
  • However, the business has become a less important part of IBM’s operations in recent years, as the company has expanded its software and services activities.
Comments (10)
  • Esekla
    , contributor
    Comments (2291) | Send Message
     
    Yeah this would be pretty surprising. However, if IBM can take the next step and move the board-level photonics highlighted in this article

     

    http://seekingalpha.co...

     

    down to the chip level then maybe the current silicon operations aren't going to stay relevant.
    7 Feb, 07:48 AM Reply Like
  • aus_tech_community
    , contributor
    Comment (1) | Send Message
     
    Board level photonics are still far from mainstream. The cost of the technology is what currently prevents this idea from making its way to product. In addition, the transistor density offered by current silicon technology is orders of magnitude higher than anything achievable from photonic computing.

     

    IBM seems hesitant to spend money on anything the doesn't have sky-high returns. I seriously doubt they would be interested spending money to get bleeding edge technology into the mainstream.
    7 Feb, 05:31 PM Reply Like
  • Vacated
    , contributor
    Comments (25) | Send Message
     
    Steve Zolman was right, it's turning into a big garage sale. PC unit, POS unit, x86 unit, chip fab unit, and eventually Services unit. I can't help but envision the inside of the Armonk facility, a bunch of fat lazy execs loading money in their duffle bags right before turning out the lights. Sad!
    http://bit.ly/1duX95m
    7 Feb, 07:52 AM Reply Like
  • jhesr
    , contributor
    Comments (350) | Send Message
     
    This is probably just for the manufacturing plants. I'm positive IBM would keep the power processor design and research. They would just be going fabless which is what most companies do already. Just like Qualcom, Nvidia, Apple, AMD, etc are all fabless.

     

    It costs a fortune to keep the manufacturing up to date. Much better to use Taiwan Semiconductor, Global Foundries, etc for the manufacturing.
    7 Feb, 10:46 AM Reply Like
  • Lowcarb
    , contributor
    Comments (48) | Send Message
     
    jhesr,
    There is a reason why IBM has owned it's own fabs. They get to tweak and tune the semiconductor technology to fit the needs of the server to get maximum performance for their processor design. When you own the fab you can get the technology tuned to match your needs. When you are a customer of a foundry you have to tweak your design to perform as best as it can with the technology performance that the foundry delivers.
    Yes, you can pay for a level of line tailoring but not near the magnitude of what IBM can do for themselves in their own fab.
    By owning the semiconductor fab IBM can also do whatever is needed to deliver parts to ship a new mainframe offering. If you are buying from a foundry there is only so much expediting that you can buy, and you will probably pay dearly for it.
    If you are using the same semiconductor technology and foundry that your competitors are using how do you differentiate your product?
    If IBM sells off the Microelectronics unit you can bet that in a generation or two there won't be anything special enough about a Z system that makes a customer want to stay with IBM.
    8 Feb, 09:57 PM Reply Like
  • alpine
    , contributor
    Comments (880) | Send Message
     
    I think the potential acquirer, given everything that is going on in the chip business these days, could well be a Samsung, and just as IBM helped Legend Holdings morph into Lenovo with its iPad sale, it could sell to fledgling the ultimate China Chip Inc., i.e SMI (Semiconductor Manufacturing International Corp). This way, China Inc will crawl its way into so many of the mainstream IT businesses worldwide, esp. since SMI is already an ARM licencee.

     

    Something that bears watching closely, not because of what effect it has on IBM, but on how it would change the chip industry competitiveness landscape permanently, including the entire array of large chip equiment manuacturers, esp the many US based ones.
    9 Feb, 03:20 AM Reply Like
  • jhesr
    , contributor
    Comments (350) | Send Message
     
    Lowcarb, I hear what your saying but I don't think IBM will shoot themselves in the foot. Whoever they sell the plants too, I'm sure they will have an iron clad agreement in there such that IBM will not have any trouble getting them to make their chips for them.

     

    Even if IBM did not sell off the manufacturing I would think in 10+ years that the power servers will be dead by then. Its already happening, huge declines for everyone still in unix. The mainframes, i.e. system z will probably still be ok though. For system z, I think even if the power processor is no longer anything special, people will still be buying system z because of all the other things built into that system. Its not just the processors that make the mainframes so fault tolerant.

     

    I think once IBM sheds enough of their hardware business we will finally start to see top line growth again, albeit slow growth. I don't think IBM's heart is in the hardware business anymore. Its not really who they want to be and just looking at where their revenue and profits come from its already the reality. I think them shedding off most of the hardware business will be a good thing. Right now its a distraction with the constantly eroding revenue and market share. The revenue loss is keeping IBM's top line from growing which is holding down the share price.
    10 Feb, 01:17 PM Reply Like
  • mnkhorrami
    , contributor
    Comments (11) | Send Message
     
    IBM is cutting their arms and legs to feed the 2015 roadmap and $20 EPS.
    The 2020 EPS is going to be $5.
    7 Feb, 12:22 PM Reply Like
  • 1GreatCFA
    , contributor
    Comments (830) | Send Message
     
    Friend of mine. BIG TIME tech head. Told me to drop IBM like a bad habit. Said Nokia is where it's at. IBM was the place to be 20 years ago. The replacement cycle for tech [me thinks] is a lot longer than what people think.
    7 Feb, 10:54 PM Reply Like
  • chrisgar
    , contributor
    Comments (42) | Send Message
     
    Nokia ?????? Really ??

     

    9 Feb, 12:17 PM Reply Like
DJIA (DIA) S&P 500 (SPY)
ETF Tools
Find the right ETFs for your portfolio:
Seeking Alpha's new ETF Hub
ETF Investment Guide:
Table of Contents | One Page Summary
Read about different ETF Asset Classes:
ETF Selector

Next headline on your portfolio:

|