back to article IBM won't grow, says analyst firm while eyeing flatlining share price

IBM’s profits are in a “likely irreversible structural decline”, according to a house of analysts that isn't convinced the company is in a position to grow anytime soon. "We love this 100+ year-old American icon, and support management's turnaround strategy, but six+ years in, the company's recovery remains protracted and …

  1. John Savard Silver badge

    Good News

    This may be bad news for IBM, but ultimately it could be good news for everyone else.

    At some point, potential revenue from mainframes will shrink enough so that IBM will lose its fear of cannibalizing those revenues. Then, they might make the System/360 architecture, in its modern form of the z/Architecture, available in desktop and laptop computers.

    1. Anonymous Coward
      Anonymous Coward

      Re: Good News

      This may be bad news for IBM

      Only if the Wall Street scribblers are correct. As a general rule, they're pack animals, independent thought is unwelcome, and they have a bad track record of spectacularly failing to foresee the future.

      The Reg is happily towing the line here that IBM is doomed. The doom merchants may be correct (I should point out its usually me that is merchandising doom). But in this case I'm one of the (apparent) minority that are rather hesitant to write off a company that over the past year or so has had revenues of around $80 billion, has a gross margin approaching 50%, generated over $12 billion in free cash flow, has almost a further $12 billion in cash (not to mention owning one of the strongest brands in the business world, and an impressive collection of IP assets and R&D capability).

    2. Nate Amsden Silver badge

      Re: Good News

      am confused - why would anyone care to run that architecture on a desktop or laptop? I think even PowerPC which should be much easier to manage doesn't appear to have a laptop(mac of course I think was the only PPC laptop) or desktop version for many years now.

      if anything I'd expect an emulator to be made(looks like there is at least one already - ) for those that need to develop/test software on x86 systems.

  2. Bob Vistakin
    IT Angle

    This "IBM Cloud" you speak of

    When it launches, please report on it. Everyone will be curious about it in a dancing bear kind of way.

    1. Jay Lenovo Silver badge

      Re: This "IBM Cloud" you speak of

      Could always rebrand themselves IBC (International Business Cloud).

      It made some noise for IHOP...IHOB.

      Currently though, a burger from IHOP looks to be a better choice than a cloud from IBM.

      1. Joseph Haig

        Re: This "IBM Cloud" you speak of

        Could always rebrand themselves IBC (International Business Cloud).

        I realise that you are making a joke but IBM have lasted for 100 years by continually re-inventing themselves as the market changes whilst keeping a name that is sufficiently broad to encompass any type of 'business machine'. At the moment they are surviving because enough people still trust the name but I don't see their reputation lasting for much longer if they cannot sort themselves out soon. A rebrand (which I think is unlikely) would be a sure sign that they are desperate and all else has failed.

  3. a_yank_lurker Silver badge

    Not Surprised

    Itsy Bitsy Morons has been run by PHBs who have no idea what strategy means or is. They are good at mouthing platitudes and placing follow the pack not the leader. Late to major markets, unwilling to cannibalize their own legacy divisions, and general cluelessness have their traits for awhile now. At best they will become a smaller also-ran that ekes out profits in a few legacy niches and is a trivial player in the major markets.

    1. Bob Vistakin

      Re: Not Surprised

      The process those spreadsheet wielding PHB's use to manage the techies who actually do the work is known as the Intense Bowel Movement.

  4. Anonymous Coward
    Anonymous Coward

    It's no suprise that sales of enterprise software of the kind IBM sells are flat in developed economies, but there must be potential to sell their systems in developing countries (think International), or are most of those markets also mature now?

    Having an established product that just requires localisation is a huge advantage over any local startup offering. It's hard to see how they could fail - do they price themselves out of the market or are there difficult legal issues? Perhaps they should spin off a partner business to sell IBM-based solutions, tailored to the lower-cost conventions of growing markets?

    1. Yes Me Silver badge

      ...developing countries

      "there must be potential to sell their systems in developing countries"

      Nah. Their price point is way too high. If I was packaging for a developing country, I'd start with a Pi and open source code, and work down. IBM doesn't have that in its genes.

      IBM's problem for 25 years has been too low revenue per employee. You can't fix that by applying more managers to the problem. IBM's main business skill for 10 years has been "resource actions."

      All bad.

  5. Robert D Bank

    The majority of major financial institutions by far still rely on zSeries as their core platform because of its reliability, security, stability, efficiency and yes, flexibility. There's no other platform that can match it for running general workload continuously without fail. Every physical and OS layer is virtualised and has been for decades.

    Most of the comments here are lazy generalisations based on no knowledge.

    There are emulator versions of such as zDT and RD&T which work fine, but they're limited performance wise by the underlying x86 architecture. Better to run virtualised zOS images or zLinux under VM on zArchitecture.

    1. Paul Crawford Silver badge

      All of that is probably true but it is still not changing the fact that mainframe use is in slow decline. How many new or expanding businesses are thinking "You know what, lets migrate from Linux/Windows servers to a zSeries mainframe?"

      1. FozzyBear Silver badge

        Or Cloud, We really should be doing it in the cloud.

    2. James Anderson

      The mainframe is the problem.

      Not because they are bad, rather, because they are very very good, and very very profitable, unlike the rest of IBMs product lines.

      The problem is that 20 years ago the all the Fortune 500 companies would be substantial mainframe customers. Since then NONE of the new entries to the list (think Amazon, UBER, etc. ) use an IBM mainframe. Those companies with mainframes are running old ( and probably excellent ) software, but, I have not seen any major mainframe projects for about 10 years. Nobody is developing new software for these beasts.

      While the hardware is actually quite competitive the software license fees are eye wateringly expensive, and, you need a long list just to get the system running.

  6. PassingStrange

    Nothing new; IBM's been moribund since the 90s

    Despite a plethora of genuine talent at the professional level, IBM management has been largely relying on old, cash-cow revenue streams and auto-cannibalism for the last 30 years - as anyone who has actually worked for the company would have had to be blinkered not to notice.

    All that's happening is that it's reaching the point where it's sufficiently obvious that even the most disinterested investors are finally beginning to catch on.

  7. Anonymous Coward
    Anonymous Coward

    Broken off

    > IBM has the appearance of a business that may well be broken up - in the same way HP was - with the growth areas lumped together and a new home found for the less desirable bits.

    Large sections of IBM have already been sold off. For example, to Lenovo a few years ago, who appear to have done fairly well out of it.

  8. Bamboozled

    Re: Downward spiral?

    They killed their GTS/GBS market with watering down their top notch skills base with their GD policy to "save' money. The customers have found them out, who will pay top $ for a premium outsourcing support solution when you get monkeys for your money instead?

  9. FozzyBear Silver badge

    "These sentiments were previously voiced by Wall Street analyst Bernstein, who also criticised IBM's six-year share buyback plan, claiming it should have spent that $92bn on strategic buys instead of helping to prop up the value of the stock."

    But, But I wouldn't have got my yearly bonus whined another over paid preening peacock (aka CxO)

  10. trevorde


    will still get her bonus

  11. Robert D Bank

    It is true most modern startups are not moving to mainframe because they're usually developed on a tight budget and by people that have been taught a mainframe can't do what an x86 server can do, and more, without having acres of racked servers. Once it reaches a certain point it's hard to switch to a new platform.

    With increased costs for energy to run the machines and aircon etc and floor space though the cost differential is a lot closer than it was. A mainframe equivalent in size to a single rack can run zOS, Linux native, hundreds of Linux and zOS under VM, AIX and more, reliably, with unsurpassed workload management that allows it to run at 100% utilisation continuously. You now have pervasive encryption of all data including in-flight data if you choose to. Cloud is included as part of the platform as well. And all of this can be in an established and integrated Disaster Recovery process that can switch sites almost instantly. Active/Active sites are now also closer to reality.

    There's still plenty you can moan about with mainframes of course, but from what I read on these pages that applies to any of the other platforms equally. None of them do everything 100% as you may like, but you have to decide what compromises you want to make and choose. Given the demise of the mainframe has been predicted for over 40 years now, and given just how much it's advanced I think it'll be quite some time before it disappears, if at all. If only one thing survives from IBM the mainframe and it's top quality tech and research people deserve to be it.

    1. Lusty Silver badge

      I can't tell what your argument for moving to Mainframe is though? Lintel does everything you describe there and it does it in a cheaper, more scale out manner. It also exists as a real cloud offering, whereas you just used the term cloud on one rack of hardware that you have to buy (Thanks to IBM Marketing, no doubt!). Cloud is a very well defined term these days, and IBM do not participate in that industry in any meaningful way, especially not with Mainframe hardware.

      Your concept of moving to Mainframe once you start to scale is extremely out of date too. Mainframe was designed as a scale up solution which had very real limits. Modern cloud architectures are designed to use massive scale out with multiple containers or VMs to avoid the limitations of older architectures. As a result we no longer need bigger hosts. I'm sure you could use IBM systems to host your containers, but the question would be why, given the high costs?

      If you're relying on hardware/software to keep you up and available in 2018 you've missed a lot of things at architect school. These days we plan for failure, expect failure, and plan for a retry of the process. The banks have more than proven how reliable Mainframe is in reality in a modern agile world. Reliability came from the fact that in the 1980s people weren't changing things and adding new features daily. They also weren't connected to a billion customers.

      1. Robert D Bank

        A racked z-Server hosting cloud(s) can be an on or off premise aas like any other platform and can scale out the same way. On/off premise a just choice based on your business model. You can even have both, and a mixture of platforms.

        I don't know why you think zSeries cannot host multiple VM's or containers because they can. Virtualisation in the mainframe is nothing new, it's matured over decades at physical and software level. Linux can run natively or under VM, as can zOS, at the same time if you wish using ultra-fast interconnect within the processor to talk to each other. All this with 100% pervasive encryption, include in-flight. You seem to think z-Arcitecture is out dated, but I'd say it is your understanding of it that is out of date.

        Your point about a billion customers is moot as I suspect there are still relatively few companies servicing that number and it would be a challenge for any platform. And it's not just customer numbers that is relevant, it's the transaction rate that counts, along with network bandwidth, and how many applications interact to name a few.

        In terms of agility that has less to do with the platform and everything to do with the company structure and culture, standards, and available talent.

        I am not an architect, but I have had to suffer the whims of architects who too often work at a theoretical level and have no practical operational experience.

        1. Lusty Silver badge

          This is kind of the point, you think architectures struggle to scale to a billion customers, but modern cloud architectures don't struggle at all. I never said it couldn't virtualise, I said I don't see why I would run containers or Linux VMs on such over priced hardware. The only reason I can think of is clock speeds being higher, but it's very rare to have something so single threaded and latency sensitive these days. My point abou the architecture is that it is designed for scale-up solutions, and hence costs a ton of money. For scale out you're far better off using lots of very cheap systems.

          On the point about cloud - it has nothing at all to do with on prem vs off prem. Cloud is a very specific solution and Mainframe is not it whether they've slapped that badge on in the software or not. It just doesn't meet the requirements as set out by NIST, and even IBM's "cloud" isn't a cloud solution, it's just rebadged colo for the most part. All of the platforms I work with also have end to end encryption, this isn't a new thing. We also have homomorphic encryption too so that data can be used while encrypted in a database more usefully.

          I agree, you can be agile on a Mainframe. It's just that nobody is, and every change costs millions. Of course, that can be solved with Linux VMs, but then why have a Mainframe? Every argument comes down to why would you choose to go to the more expensive hardware to implement the same solution, and there are no good answers to that other than 1. you work for IBM or 2. your knowledge is decades out of date.

          1. Robert D Bank

            I'm not an IBM employee, that's just a cheap shot that takes away from anything else you have to say. At worst you might call me a Z evangelist. You could turn the same argument on yourself.

            My knowledge is lacking especially about Cloud, happy to admit that, but it's an interest. And yours about zArchitecture?

            Reading Wikipedia and the the definition of Cloud by NIST all the 5 essential characteristics are met by z-Architecture, especially the resource pooling and resource elasticity. And the energy and floor space efficiencies afforded by z-Architecture certainly counter the 'lots of cheap machines' argument. Interestingly cost is not in the list, although obviously is a major factor.

            The cost of the architecture is a big issue I agree, and that is definitely an area IBM will have to become more competitive. Whether they succeed or not will become self evident in time. It is always better to have more players in any market.


  12. Robert D Bank

    little bit of insight here

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

Biting the hand that feeds IT © 1998–2019