Comments Locked

36 Comments

Back to Article

  • rstuart - Wednesday, July 22, 2015 - link

    One bad mistake when designing a memory controller, and it all turns to shit. Life is tough when you are an engineer.

    The response to the financial markets made me laugh. The problem lies within the Engineering team. Or probably more correctly lay within the engineering team, as there are so many bloodied noses I'm sure the lesson has been rammed home and it won't be forgotten for a long time. Yet did what did the CEO say?

    > reviewing their corporate structure, capital return opportunities, and other alternatives with the goal of creating stockholder value ... have reaffirmed their commitment to return significant capital to shareholders, with a minimum of 75% of free cash flow being returned through dividends and share repurchases

    Right. That's a sure certain to fix.
  • ikjadoon - Wednesday, July 22, 2015 - link

    "The problem lies within the Engineering team".

    We have this discussion every time some company makes a foul up. Was it management pressuring engineers to meet "market demands" or engineers fouling up?

    Please, everyone, continue to armchair-speculate what *really* happened...
  • edzieba - Thursday, July 23, 2015 - link

    Given the move to 64bit cores was for a spec sheet feature rather than actual functionality, it's not hard to guess where the push for it came from.
  • AndrewJacksonZA - Thursday, July 23, 2015 - link

    +1 ikjadoon
  • Samus - Thursday, July 23, 2015 - link

    The management and marketing team screwed up by not properly evaluating market demands or predicting the next "trend"

    If the engineering team wasn't instructed to make a 64-bit chip, how is it their fault? If the engineering team made a faulty 64-bit chip, then it would be their fault.
  • icrf - Wednesday, July 22, 2015 - link

    This is what I hate about the stock market. Vocal shareholders freak out after a quarter or two of issues. Lay off what gives the company value six months before seeing how well they did and if you need them, and take the pile of money the company has and put it anywhere except into the company (ie, shareholders).

    I don't want to own a stock for six months. I want to own it for six years, and I'm okay with the value tanking along the journey, just make the company better in the process and pay me off when times are good. I'm worried they could be the Intel of the ARM space, but short sighted moves like this will kill them.

    If, on the other hand, they over expanded when times were good and this is just an adjustment back to what's actually needed, then that's another matter entirely. It doesn't sound like that, though.
  • iWatchHogwash2 - Thursday, July 23, 2015 - link

    That's a thoughtful observation, good one.
  • Nagorak - Thursday, July 23, 2015 - link

    On top of that, the idea of breaking up the company is incredibly stupid. You take a profitable, but erratic chip making unit and put it on its own and what is going to happen? All it takes is a couple blunders and they go the way of AMD. The more stable income componet helps maintain the chip making unit during tough times, so it can turn things around.
  • Samus - Thursday, July 23, 2015 - link

    I agree, Qualcomm's "risky" chip business needs the stable income of the licensing business in order to weather the bad times.

    What the hell ever happened to the concept of corporate diversification?
  • Ananke - Thursday, July 23, 2015 - link

    Corporate diversification is done through capital portfolio diversification, i.e. the shareholders will have investment in chip making and licensing only corporation, and will sell one of these if they don't perform.
    Unfortunately, in the long run this is not competitive strategy against Chinese government backed companies or vertical integrated Koreans like Samsung.
    RIP Qualcomm...and all fabless businesses actually. Fabless is not competitive globally.
  • ppi - Thursday, July 23, 2015 - link

    Yeah, diversifiaction on shareholder level is the theory ... but say if you close down chipmaking business, because it did not perform well enough for two quarters, you cannot just reopen it instantly, it's gone.

    Also, for IP licensing, you need to continuously produce something to license, and not just sit back. And that stuff has very good synergy with chipmaking.

    (at the same time, it needs to be said, that conglomerates that have completely divergent business)
  • jjj - Wednesday, July 22, 2015 - link

    The outlook for MSM at 170-190 million units was shocking and makes one wonder.
    They shipped 225 million in Q2 (calendar) and for Q3 they project just 170-190 (so midpoint 180). Last year Q2 was 225 and Q3 236 million units. So this year Q3 is 66 million units bellow last year Q3, while Q2 were equal. At the same time the smartphone market grew. They lost the Galaxy S6 and very likely the soon to launch Note 5 but in the end that's not such a huge volume per quarter . At the same time Apple grew very nicely on year and Apple's share gain is also Qualcomm's share gain.
    Qualcomm claims that it's the S6, inventory glut and poor sales for SD810 but all that seems far too little for such a huge drop.
    I don't want to start rumors and this is not a fact or something that others might agree with, it's just the only reasonable conclusion i am able to arrive to. Given the numbers i am convinced that they lost the modem in the next iphone. Would be very unexpected and maybe it's not the case but it sure looks like it., the drop is just huge - again don't take it as a fact, it's just my suspicion and you are free to arrive to your own conclusion.
    Sadly that kind of outlook also makes it likely that SD620 won't ship in high volumes in Q3. Kinda hoped we see it soon in devices.
  • jjj - Wednesday, July 22, 2015 - link

    "So this year Q3 is 66 million units bellow last year Q3"

    56 million not 66* - no edit but figure the type is too big not to fix it.
  • Gunbuster - Wednesday, July 22, 2015 - link

    Didn't help that HTC made a crappy SD810 phone and Microsoft forgot to build one altogether...
  • Impulses - Thursday, July 23, 2015 - link

    Like HTC's numbers would really mean much... At the high end, losing the SGS6 probably meant more than any other design wins/losses put together. You gotta figure Samsung wasn't gonna choose Qualcomm indefinitely tho, even if Qualcomm didn't fall behind.
  • lmcd - Thursday, July 23, 2015 - link

    Forgot? Why would they put in the effort? The 810 wasn't that great of a chip. I'd rather an 805 than an 810.
  • Michael Bay - Thursday, July 23, 2015 - link

    Considering that MS is using <i>only</i> Qualcomm SoCs in their phones, it`s more like they saw the heating issues and just didn`t bother.
    I wonder if their new phones will suddenly feature Atom.
  • jjj - Thursday, July 23, 2015 - link

    They have a handful of devices in retail with SD810/808. LG Flex 2, HTC, Xiaomi, ZTE , LeTV , Sony and the G4 with SD808. Those are about to double with the ZTE Axon that just launched now in Q3, the Oneplus 2, ZUK Z1 (new Lenovo brand), a phone by Coolpad and Qihoo 360 with SD808 a likely Lenovo X3 with SD808 plus looks like Moto and Microsoft have some devices with SD810/808.
    So there is no way 810 inventory can be all that much and they are doubling the number of devices with 810/808 in retail during Q3. Something else has to be off in a big way when it comes to units.
  • jerrylzy - Friday, July 24, 2015 - link

    It's unlikely that Apple chose not to use Qualcomm's modem for iPhone 6S.
  • karthik.hegde - Thursday, July 23, 2015 - link

    Major difficulty for Qcom is their huge dependancy on Apple and Samsung. If the rumour about Apple using Intel modem in their next Iphone and Samsung developing its own solutions, then Qualcomm is in problem. They'll have to survive with royalty only.

    On the processor side, MediaTek, using the off the shelf ARM CPUs, GPUs and other IPs is pushing hard.

    Will be an interesting battle in ARM field, too bad this is not the case in x86.
  • coburn_c - Thursday, July 23, 2015 - link

    Told them there was no reason to go 64-bit yet. Told them it would suck. All these faggots here said I was stupid, but I could have saved them.
  • exmachiner - Thursday, July 23, 2015 - link

    Qualcomm didn't decide to move to 64-bit. Their hand was forced. It is clear now that they had estimated the 64-bit transition wouldn't happen until 2016-17 (when smartphones would actually ship with 4GB RAM).
  • Samus - Thursday, July 23, 2015 - link

    That was an epic marketing evaluation screw-up. History has a habit of repeating itself: AMD introduced 64-bit architecture to the desktop nearly decade before 4GB of RAM was standard.

    The difference is 64-bit x86 was introduced by an underdog, so it wasn't taken as seriously as when, say, Apple, introduces the technology. Intel was able to counter AMD with a 64-bit x86 ISA within two years. Qualcomm is, for the most part, doing the same thing. But as the post and others have mentioned, the ARM industry is interestingly different than x86 because there are so many players.
  • Klinky1984 - Saturday, July 25, 2015 - link

    Intel already had IA-64 available well before AMD introduced AMD64(x86-64), but no one wanted Itanium processors except for certain niches due to compatibility problems. People(Microsoft) did take AMD64 very seriously, and essentially forced Intel into implementing their own version(EM64T/Intel 64).

    Right now Qualcomm is more like Intel, with it's heat issues(Pentium D) and failure to transition to 64-bit computing(IA-64). Intel was big enough to weather the storm. Qualcomm probably not so much.
  • Kvaern2 - Thursday, July 23, 2015 - link

    What I gather from your post is that out of the all Arm V8 improvements the only word you understood is 64 bit?

    I'm surprised they didn't listen to you.
  • coburn_c - Saturday, July 25, 2015 - link

    If I understand correctly, had they stayed on armv7 there chips would be faster, cooler, and with better battery life. I know they would lose out on ALLL THOSE 64-bit (and... other improvements?) only applications but...
  • sseemaku - Thursday, July 23, 2015 - link

    These investors never have patience. Seems like you cannot invest in anything long term when your is a public company.

    That said, I didn't think it would take so long for qualcom to bring out ARM V8 based SOC. Seems like they had a different roadmap!
  • name99 - Thursday, July 23, 2015 - link

    Why are you blaming the INVESTORS here? It was QUALCOMM that did not invest in an ARMv8 core years ago.
    The investors are responding more or less rationally to the fact that the entire management complex, from board to CEO, seem not to have a realistic plan for the future.
  • Shadowmaster625 - Thursday, July 23, 2015 - link

    I'm glad they were punished for an early attempt at providing useless and pointless 64 bit support. There is no reason to rush with it, especially when it costs you half your profits. It will still be another 3+ years before devices even begin to need 64 bit. Hell, it is rare to even find a phone that has 4GB of RAM. There was no reason to try and push it out before its time. If they had built the best possible 32 bit SoC they could have, they would have made a lot more money. Instead, they clobbered themselves with the stupid stick.
  • K_Space - Thursday, July 23, 2015 - link

    I think they're being punished for their reactive move rather than the inclusion of 64 bit support per se. I still recall when the A7 launched how well it was received. The v8 instruction set featured heavily if I remember correctly. Apple had a record Q4 2013 (admittedly not entirely due to the A7, but 2 new iPhones, iOS 7, new iPad mini with Retina Display, iPad Air, etc).
    My point being, from a non-technical investor POV A7 succeed because it ticked the 64 bit feature box FIRST (OK that's a bit simplistic, it was an excellent chip and a first for Apple); 810 was late to the party (and had a hot head). I'd image it'd have been a different story if 808/810 launched before the A7 (or equivalent).
  • name99 - Thursday, July 23, 2015 - link

    Let's recall some of your earlier predictions. In 2011 you were telling us that tablets were a fad. You were likewise criticizing the "icrap 3GS". Then in 2013 you were telling us how the smartWatch was a stupid idea.

    Might I suggest that you are basically EXTREMELY conservative by temperament, and that that's not a useful trait in someone trying to predict the tech future, because you don't have the inclination or the temperament to look beyond what's immediately visible?
    You see an early iPad, an early iPhone, an early smartWatch (you were criticizing the Galaxy Gear I think in 2013, but I'm sure you have even more vitriol for the apple Watch today) and all you see is a a tablet is lousy at doing laptop jobs; a phone has a screen that's too small and a horrible typing experience; now a watch can't do the things a phone can do. You see ONLY the negatives.
    Other people (and it turns out, these are the people with substantially better tech prediction track records) see the possibilities. Not just that the tech will get better --- the CPUs will get faster, the batteries will last longer --- but that these devices will be used in whole new ways that just don't map onto how you used the previous device.

    I'm laying this all out because I think this is exactly the same problem that has crippled QC (and to some extent MS). They look at the world as it is today, and consider only the jobs that are being done today, and how those jobs are best served. And so yes, IF your vision of a smartphone is that ALL IT IS is basically a 2007 sort of phone with a little bit of Java, a media player, a crappy web browser, then sure staying at 32-bits for years is fine. But if you look at a smartphone and ask "what could I do with a kick-ass pocket computer?" then you put yourself on a very different trajectory.

    MS had much the same blindness (certainly under Ballmer, maybe still under Nadella). They could kinda see that smartphones/pocket computers were important, but only at an intellectual level. They didn't really get the excitement, the thrill, the feeling that "*I* want one, because these are all the things I want to do with it". And so they don't ship any sort of leading edge phone --- all they ship is me-too phones that are just going through the motions. Their one successful mobile product, Surface Pro, is basically a way to keep going with the PC dream of 20 years ago, without having to imagine anything new.

    Relevance to Apple Watch and the way it is viewed by different segments is left as an exercise to the reader. But before deciding who to listen to, ask yourself which of these different viewpoints has had the more accurate track record over the past ten years or so.
  • K_Space - Friday, July 24, 2015 - link

    You are not referring to Brett? He started for Anandtech in 2014, no?
  • Michael Bay - Friday, July 24, 2015 - link

    Your "kick-ass pocket computer" is just that - a phone with tacked on bullshit of varying degrees of usefullness. It still has a crappy web browser, a bit of java and so-so media player. Better yet, it`s completely ubiquitous by now, with only freaks getting excited over a piece of plastic.

    iPad and tablet at large is over, smartphone is over, and smartwatch won`t ever begin.
  • tipoo - Thursday, July 23, 2015 - link

    I hope this is lesson learned and Qualcomm stops making small, low IPC, high clocked cores. Clearly not the way to go, least of all on mobile.
  • name99 - Thursday, July 23, 2015 - link

    "It’s not fair to blame all of Qualcomm’s woes on one issue, but Qualcomm’s issues began when Apple made the surprise move to 64-bit."

    You can call it "one issue" but QC seems to have suffered from industrial strength complacency in 2013. The problem is not that Apple released the A7 (a CPU which doesn't even compete with QC for any sales),, it's that QC seems to have been utterly blind to the idea that CPUs were improving, and to imagine that they controlled the high-end ARM market.
    You have to be crazy delusional not to be aware that 64-bit is coming, that other companies can design CPUs; but QC apparently had NO plan B. They surely knew about the existence of the ARMv8 ISA, maybe they even knew of the existence of the A57, but in spite of that, no planning?
    WTF???

    Even nVidia, the gang that can't shoot straight in the ARM CPU space, have managed to put together a credible ARM-64 CPU faster than QC. The single most important job of a CEO is strategy planning .

    To be fair, maybe this is Paul Jacobs fault (and the reason he was forced out as CEO in 2013)? So Steve Mollenkopf is the Satya Nadella, trying to clean up after his mess. Of course, like Nadella, we've yet to see if he has a reasonable vision for QC's future once he's cleaned up.
  • jerrylzy - Friday, July 24, 2015 - link

    Actually they designed a very fast 32bit chip called "Taipan," but it was soon abandoned after Apple released Cyclone.

Log in

Don't have an account? Sign up now