back to article Intel tried selling software before. Will it succeed this time?

Intel is in the midst of a massive effort to rekindle its chipmaking machine, in addition to building a business out of selling software for the second time in less than a decade. For the past several months, Intel CEO Pat Gelsinger has not been shy about the company's software revenue ambitions. The CEO reportedly lamented in …

  1. Joe Dietz

    As a software engineer, never work for a hardware company (and vs. versa no doubt)

    I was at McAfee during the Intel era... It was always a complete mystery to everybody wtf it meant to put 200+ MB of MD5 hashes 'into the silicon'. Though it was apparently on several executive bonus plans that we needed to increase sales of 'sockets', not that anybody in McAfee mgmt chain had a clue how to make that happen. I'm pretty sure the confusion on the Intel side was pretty much the same.

    See 'security'... as Intel understood it meant door locks, 'security' as McAfee understood it meant 'background checks'. I could see a sort of world where those two notions worked together... but customers where not that interested in paying _more_ for the privilege and given how Intel slices up their CPU features in pretty arbitrary ways... you end up with the 'why does this brand-new computer not run windows 11' problem all over, or in this case 'why does my door-lock fail open'. Thus ended any actual engineering attempts to bring those ideas together.

    I'm not sure much has changed, but the main thing is software businesses should be software businesses.

    1. OhForF'
      Trollface

      sell more sockets

      >we needed to increase sales of 'sockets', not that anybody in McAfee mgmt chain had a clue how to make that happen<

      So the virus scanner hogging all available CPU resources was an attempt to get the users to buy more sockets?

      1. Yet Another Anonymous coward Silver badge

        Re: sell more sockets

        No there is a shortage of CPUs and Intel have problems producing them

        While sockets are relatively cheap and easy to make and have quite high profit margins.

        The license now simply requires 186 pins / seat

  2. katrinab Silver badge
    Windows

    Aren’t they doing it back to front?

    Apple does hardware and software, but their starting point is an end-user product that does something, then they design the hardware and software to make it do it. I know they are not universally loved, but there are lots of people out there who like what they do.

    Intel’s starting point seems to be, “we need to find a use case for this chip we designed”.

    1. Yet Another Anonymous coward Silver badge

      re: Intel’s starting point seems to be, “we need to find a use case for this chip we designed”.

      I think that suggests an optimistic level of strategic planning:

      Dept 1, We have some existing chips that we can make

      Dept 2, We have an idea of how to make a new chip

      Dept 3, We can't make the new chip but we aren't going to tell Dept 2 cos they report to a different VC

      Dept 4, We seem to have accidentally acquired a bunch of software companies - and my bonus relies on us having a plausible plan for making money from them

      1. Youngone Silver badge

        Re: re: Intel’s starting point seems to be, “we need to find a use case for this chip we designed”.

        That is hilarious and sad at the same time, and you've also described two of my former employers.

    2. EnviableOne Silver badge

      that was Jobs, and how he worked.

      Apple is now basically a marketing machine that makes it look like their next minor innovation is gods gift to technology, even though Samsung, Sony or google did it 5 yrs ago

  3. tiggity Silver badge

    software to show off the hardware

    A long time ago* I used Intel C/C++ compiler (on Windows) as it produced code faster than MS compiler, lower memory use and less memory read / writes & it was also easy to exploit GPU for extra processing (back in the days where palming off work to the GPU could be a PITA to code).

    Whether it "showed off the hardware", or merely good compiler writers who can say, but it left a favourable impression.

    That's the sort of thing I might expect a "hardware" company to do, "near the metal". so complier, VM code to optimize interaction with the hardware etc all seems the stiff I would expect in terms of software.

    AV stuff less so, as a big part of that is OS specific (though IME is obviously a nice juicy non OS target & has had exploits created over the years)

    * Lots of real time image processing and performance was of key importance hence various compilers evaluated & Intel were best at the time for the processing we had to do.

    1. gerryg

      Re: software to show off the hardware

      I understand that Intel have the best Linux distro but don't distribute it.

      If true, it would be interesting to understand why IBM/RedHat works but Intel/Linux distro isn't an option

  4. iron Silver badge

    > you've got McAfee. They've got some great stuff.

    Said no IT person ever.

    1. Yet Another Anonymous coward Silver badge

      > you've got McAfee. They've got some great stuff.

      Personally he seemed to have some pretty great "stuff"

  5. msobkow Silver badge

    I read that as "after unloading some Windows-only properties to focus on cross-platform and cloud markets"...

  6. martinusher Silver badge

    Too much financial shenanigans, not enough engineering?

    Probably the worst thing that ever happened to Intel was the near monopoly they had for PC processors. This didn't so much turn on the money taps as open up a financial fire hose which completely skewed the company's priorities. Instead of being primarily a chip maker it became dedicated to enhancing and preserving that monopoly by any means and at any cost. The problem is that eventually WinTel will run its course....and then what?

    The early Intel sold the complete range, everything needed to build systems with their processors. Not everything was perfect -- their MDS was ridiculously overpriced, for example -- but they had everything plus a range of parts to suit different applications and price ranges. Gradually they dumped stuff, I'd guess because compared to the PC business the revenue was insignificant and profit contribution negligible. the problem with this approach wasn't just WinTel -- it became increasingly inconvenient to build anything other than a PC with Intel parts. This left a huge market for the taking, a market that they then discover can do the kinds of stylized application frameworks just as well as the traditional WinTel PC can.

    Success also had another drawback. Intel, being an early entrant to the LSI semiconductor business, had its own fab methodology which over the years became theirs and theirs alone. It basically became "Intel" and "Everyone Else". Eventually "Everyone Else" outstripped them and the cost of switching over to them (and the huge NIH drag) meant it would take a lot of time, effort and money.

    I daresay that they'll survive and continue to be a major player because they're so huge.

    1. Fifth Horseman

      Re: Too much financial shenanigans, not enough engineering?

      Intel were originally a memory manufacturer - DRAM and the UV EPROM are both Intel inventions. By the start of the eighties, they probably controlled fifty percent of the world memory market. Then Hitachi, NEC, Mitsubishi and Toshiba rather enthusiastically joined in the fray, and that particular gravy train ground to a rapid halt.

      In that context, the 8086 saved Intel. The iAPX432 was a dismal failure - fascinating architecture but abysmal performance - and the 8080/85 had been displaced by cheaper and faster but largely code-compatible Z80 derivatives. Certainly Intel sold at least ten 8048/51 microcontrollers for each 8086/88, but these are low price, low margin devices.

      I think things started to go awry with the Pentium. At this point Intel started to exploit its market position. Second source manufacturing licenses ended, forcing AMD down the route it took and, as you say, it started to make little sense to build anything other than PC architecture products around x86. Intel's attempts to broaden its product portfolio stalled. The i860 was good on paper but much less so in the real world, the less said about Itanium the better. The i960 is/was in fairness fine, but very much a niche market processor.

      I think Intel will survive, but it will definitely have to change focus. Their semiconductor processes may be a bit out of sync with everyone else but given the uncertainties and capacity issues in the chip supply chain at the moment, their fabs may well become their biggest asset. If (OK, when) the demand for x86 dries up, I am sure plenty of people will pay to use that manufacturing capacity. VHDL or Verilog are pretty much process agnostic. Intel's purchase of Altera probably won't hurt here either.

      As for software, as a few people have already pointed out, the Intel C/C++ compilers are very good...

  7. JacobZ

    Deja vu all over again

    Intel has had many forays into software - remember it's stewardship of Lustre, for example, which was decidely... uh... lackluster? And every time it fails for the same reason: ultimately, Intel cares only about selling hardware.

    And you can't make good software unless your goal is to make good software. So Intel always ends up making software that is neither good in its own right nor good at driving hardware sales. Then it loses patience, throws the software business out, and after a couple of years repeats the cycle. Prediction: DAOS will be next to be thrown out of the boat.

    1. msobkow Silver badge

      Re: Deja vu all over again

      There are exceptions, of course. Their marvellous C++ compiler. Their other libraries and toolchains and drivers.

      They do write some excellent software; the problem is they're a lousy software product steward.

  8. a_yank_lurker Silver badge

    My Doubts

    Chipzilla makes specific hardware they sell to system builders. They do not understand how the system will be used by the end user. They do not directly deal with the end user. Chipzilla only has a handful of major customers they need to pay attention.

    Software slingers usually sell to the nominal end user and thus have a much better of their needs. Also, software slingers have many more customers than a handful OEMs.

    The markets are very different and require very different approaches.

  9. Ace2 Bronze badge

    Would you buy software (from anyone) that only ran on Intel processors?

    Maybe there’s a market for that, but I’m not sure how big it would be.

    1. jonnycando

      In the middle ages…somewhere between today and the PDP-8 there was Motorola and Intel….at least it seemed that way for a while and there were Apple // and macintosh on one hand and the so called PC on the other. Not to mention commodore and so on. All the software you could buy was either for PC with Intel chips or Apple with Motorola chips. A little crossover that required complete rewriting to effect but not a lot. Even today WINDOWS will insist on being installed on PC that is either Intel or AMD or any Intel compatible system. Of course Intel might come up with software that only runs on PC but they will not be the first or last. Of course I am not the type to think up what sort of software they could do that could have broad enough appeal to make a version for every OS and chipset and CPU. Seems rather like reinventing the wheel.

  10. Anonymous Coward
    Anonymous Coward

    ~ 2000 - 2010 I was working on image processing for industrial applications - feature extraction etc - to company was using MS as the OS for some reason. Intel provides (or at least provided) a software library of math routines optimized for Intel CPU's: FFT , inverse matrix solving, etc. Great stuff.

    That kind of Intel software makes a lot of sense.

    1. Anonymous Coward
      Anonymous Coward

      "Intel provides (or at least provided) a software library of math routines optimized for Intel CPU's: FFT , inverse matrix solving, etc. Great stuff."

      Would that be the Kuch and Associates, Inc stuff? Clever stuff which used to be sold by multiple vendors as well as Kuch themselves?

      Then KAI's stuff was acquired by Intel. and vanished :(

  11. aki009

    The year the Intel died

    It's been more than 2 decades, but back in the 90's I liked going to Intel conferences.

    They were all about tech, details, more tech, cool stuff, and so on. And then something changed. From one year to next tech became a side show and it was all about marketing.

    I don't remember much about the keynote that year, but out of a gadzillion slides it seemed only 2 or 3 were about tech. The rest was fluff.

    That's when Intel died. The carcass just doesn't know it's dead yet, kind of like Nokia took years to figure out it was done in the phone business, or Russia in the superpower one.

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

Other stories you might like

Biting the hand that feeds IT © 1998–2022