Apple's PA Semi buyout motivated by assets, not products

1235

Comments

  • Reply 81 of 105
    onlookeronlooker Posts: 5,252member
    Quote:
    Originally Posted by SpamSandwich View Post


    To me, this just shows that MacWorld has no clue. Apple buys the talent and processes to create all new architectures. Why in the world would they want to hand over their innovations/breakthroughs to Intel after all is said and done. Apple wants to dominate in another area of their business and beat Intel at it's own game. That's pure Steveness. Apple moves from partner to partner like a vampire and in time, and when it no longer makes sense to deal with Intel, their bloodless husk will be cast off also.



    HIstorically speaking apple does this all the time. It started even before the acquisition of Raycer graphics. I think your being unreasonable, and irresponsible in stating things like that when there is a pattern of behavior from Apple that does exactly what this writer claims. You have nothing to back up your paranoia.
  • Reply 82 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by TenoBell View Post


    Exactly what part of that statement is assumption that cannot be backed up with information from the sources?



    Teno, it's too general. When they bought that gpu design firm a few years ago, Jobs said about the same thing, but nothing happened.



    What I think happened there was that Apple used their expertise to design better SOFTWARE that worked with gpu's, as we've seen with their core technologies. Having in-house knowledge of how these chips work at their most basic level allows much better software.



    I have some feeling that Apple may be doing some of the same here, as well as possibly wanting to do a bit of specialized x86 research.



    These guys are smart. but I don't see Apple licensing ARM to start with it all over again.
  • Reply 83 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by SpamSandwich View Post


    To me, this just shows that MacWorld has no clue. Apple buys the talent and processes to create all new architectures. Why in the world would they want to hand over their innovations/breakthroughs to Intel after all is said and done. Apple wants to dominate in another area of their business and beat Intel at it's own game. That's pure Steveness. Apple moves from partner to partner like a vampire and in time, and when it no longer makes sense to deal with Intel, their bloodless husk will be cast off also.



    There is just NO WAY that Apple is going to want to compete with Intel.



    No way, and to no purpose.



    Apple can't compete with the sales economies that Intel can deliver. That was the problem they had with IBM.



    Are they going to want to go through that all over again, except, this time, taking the monetary risk upon themselves?



    No!



    Apple does not take on that kind of risk these days. They don't need to either.
  • Reply 84 of 105
    sc_marktsc_markt Posts: 1,393member
    I read something today that said Amiga was going to use this chip. This got me thinking that maybe Microsoft was exploring making it's own computers and was looking to buy PA semi??
  • Reply 85 of 105
    hmurchisonhmurchison Posts: 12,215member
    Quote:
    Originally Posted by melgross View Post


    There is just NO WAY that Apple is going to want to compete with Intel.



    No way, and to no purpose.



    Apple can't compete with the sales economies that Intel can deliver. That was the problem they had with IBM.



    Are they going to want to go through that all over again, except, this time, taking the monetary risk upon themselves?



    No!



    Apple does not take on that kind of risk these days. They don't need to either.



    I don't quite agree. Competing against intel for Desktop/Server CPU and Chipsets is a foo's paradise but Intel is not quite the stalwart in every area. They have their hands full trying to supplant ARM at the low end power efficient arena of mobile computing. Intel hasn't had success everywhere (remember their "success" with LCoS? They bailed)



    As much as I like Apple's partnership with Intel they aren't the incumbent everywhere. I think Apple will happily maintain a healthy mix with Intel enjoying the proverbial %80 of the processor mix.



    I'd like to see what Apple can do with PA Semi IP. It's really about Apple having the ability to deliver products based on a chipset that is tailored to their needs. I have no idea what that is yet but it's easy to see how Apple can leveredge Multi-touch in many ways beyond the iPhone/iTouch.



    I'm not even so sure Apple has issues with PPC. They're working on infusing LLVM more into their development tools which target different processors with optimizations.
  • Reply 86 of 105
    splinemodelsplinemodel Posts: 7,311member
    Quote:
    Originally Posted by melgross View Post


    I don't see that anymore. Once it made sense. In fact, with the PPC alliance, it did make sense.



    But with Apple on x86 for its Macs and ATv, it makes more sense to move to x86 all 'round. Why should Apple support, real time, two chip architectures? ... I can't find a reason for this. The word is out that they don't want the products, so what else is left? Expertise? That's what the word is. But for what?



    You are too narrow-focused in your analysis.



    - P.A. Semi has expertise in building low power circuit blocks for ASICs, SoCs, and CPUs.

    - Apple has many products that use custom ASICs, SoCs, and CPUs.



    The PPC involvement, it seems, is extra. I personally think the PPC is a better chip that the x86 for use in everything other than a PC (term used loosely), but my personal opinion is probably not what drove Apple's decision to acquire PA Semi. That, it seems, was motivated by the two points above.



    I'd also like to objectively point out that the cost of producing custom silicon is not very high anymore. I really and truly don't understand your objections to Apple improving its in-house design capabilities via PA Semi -- as said, Apple already produces a handful of custom chips. Most companies in the mobile phone and embedded device spaces have fairly expansive in-house chip design capabilities.
  • Reply 87 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by sc_markt View Post


    I read something today that said Amiga was going to use this chip. This got me thinking that maybe Microsoft was exploring making it's own computers and was looking to buy PA semi??



    MS does not support PPC chips. Once upon a time, when NT first came out, it supported the PPC. But that was quickly withdrawn as MS decided to just back x86. At the time, the early '90's, it was thought that the superior PPC was going to supplant the x86, but once MD dropped that support, before it even got off the ground, that ended it.



    I doubt MS has any PPC code left that would be usable, even in the very unlikely scenario they were interested.
  • Reply 88 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by hmurchison View Post


    I don't quite agree. Competing against intel for Desktop/Server CPU and Chipsets is a foo's paradise but Intel is not quite the stalwart in every area. They have their hands full trying to supplant ARM at the low end power efficient arena of mobile computing. Intel hasn't had success everywhere (remember their "success" with LCoS? They bailed)



    As much as I like Apple's partnership with Intel they aren't the incumbent everywhere. I think Apple will happily maintain a healthy mix with Intel enjoying the proverbial %80 of the processor mix.



    I'd like to see what Apple can do with PA Semi IP. It's really about Apple having the ability to deliver products based on a chipset that is tailored to their needs. I have no idea what that is yet but it's easy to see how Apple can leveredge Multi-touch in many ways beyond the iPhone/iTouch.



    I'm not even so sure Apple has issues with PPC. They're working on infusing LLVM more into their development tools which target different processors with optimizations.



    LCoS failed with Sony as well. Intel was just smarter about getting off earlier. Wasn't Intel's fault.



    One thing I believe, when Jobs said that they were "through" with PPC, he meant it.



    I just don't see Apple taking the huge expense, and risk, associated with doing a processor chip. The costs are almost impossible when you have no history. And PA's history isn't enough. Despite the ARM knowledge that is there for a couple of people, it's fairly old knowledge now. They would have to get up to speed, and the rest of the staff would have to learn even more.



    And for what? I just don't see it.



    Apple could easily sink another $500 million into doing a new ARM version just for themselves, with no guarantee it would be better than whatever Intel will have two years from now when Apple's chip would come on line.



    I just don't see it.
  • Reply 89 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by Splinemodel View Post


    You are too narrow-focused in your analysis.



    - P.A. Semi has expertise in building low power circuit blocks for ASICs, SoCs, and CPUs.

    - Apple has many products that use custom ASICs, SoCs, and CPUs.



    The PPC involvement, it seems, is extra. I personally think the PPC is a better chip that the x86 for use in everything other than a PC (term used loosely), but my personal opinion is probably not what drove Apple's decision to acquire PA Semi. That, it seems, was motivated by the two points above.



    I'd also like to objectively point out that the cost of producing custom silicon is not very high anymore. I really and truly don't understand your objections to Apple improving its in-house design capabilities via PA Semi -- as said, Apple already produces a handful of custom chips. Most companies in the mobile phone and embedded device spaces have fairly expansive in-house chip design capabilities.



    I'm not denying that the peripheral chips they could need are possible here. But, to spend what will end up being over $300 million by the time this is done, must involve more than that, and I don't see it as being a new ARM, or esp. a new PPC.



    I agree with the Macworld article. That's the way I see it. I said it before the article came out, so I'm not parroting it.
  • Reply 90 of 105
    splinemodelsplinemodel Posts: 7,311member
    Quote:
    Originally Posted by melgross View Post


    I'm not denying that the peripheral chips they could need are possible here. But, to spend what will end up being over $300 million by the time this is done, must involve more than that, and I don't see it as being a new ARM, or esp. a new PPC..



    Low power IP is good. PA Semi basically has 150 people who have been working on the technology for several years now. $300M is very much a reasonable price for that. Not only does Apple get good IP, but by buying PA they encourage these kinds of startups. It's always easier to be a pessimist, but in this case I think you're pessimism isn't justified: Apple has plenty of money for making these kind of investments. This is a good thing for everyone, and even if it's a bust it's not a big deal for Apple.
  • Reply 91 of 105
    solipsismsolipsism Posts: 25,726member
    Quote:
    Originally Posted by Splinemodel View Post


    Low power IP is good. PA Semi basically has 150 people who have been working on the technology for several years now. $300M is very much a reasonable price for that. Not only does Apple get good IP, but by buying PA they encourage these kinds of startups. It's always easier to be a pessimist, but in this case I think you're pessimism isn't justified: Apple has plenty of money for making these kind of investments. This is a good thing for everyone, and even if it's a bust it's not a big deal for Apple.



    Where is the pessimism? That implies that the deal was a bad choice for Apple. Melgross and others have merely made pragmatic assumptions of the future use of P.A.Semi based on their current product line and Apple's current need, which are not transparently congruent at this time.
  • Reply 92 of 105
    SpamSandwichSpamSandwich Posts: 30,837member
    Quote:
    Originally Posted by onlooker View Post


    HIstorically speaking apple does this all the time. It started even before the acquisition of Raycer graphics. I think your being unreasonable, and irresponsible in stating things like that when there is a pattern of behavior from Apple that does exactly what this writer claims. You have nothing to back up your paranoia.



    As Andy Grove is oft quoted, "Only the paranoid survive".
  • Reply 93 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by Splinemodel View Post


    Low power IP is good. PA Semi basically has 150 people who have been working on the technology for several years now. $300M is very much a reasonable price for that. Not only does Apple get good IP, but by buying PA they encourage these kinds of startups. It's always easier to be a pessimist, but in this case I think you're pessimism isn't justified: Apple has plenty of money for making these kind of investments. This is a good thing for everyone, and even if it's a bust it's not a big deal for Apple.



    I'm not a pessimist. I'm just saying that I don't agree with what some are saying here as to Apple's interest. I'm not denying that Apple has a purpose for this, and it may be a good one. But I think their purpose is different from what is often being said it is, here.
  • Reply 94 of 105
    tenobelltenobell Posts: 7,014member
    Quote:

    Teno, it's too general. When they bought that gpu design firm a few years ago, Jobs said about the same thing, but nothing happened.



    There isn't much Apple can do with in house GPU technology to improve its products or their sales. This is an entirely different story for embedded CPU's.



    Quote:

    I have some feeling that Apple may be doing some of the same here, as well as possibly wanting to do a bit of specialized x86 research.



    If Apple had bought a company that specialized in designing embedded x86 chips. Then I would agree Apple will likely move all of its products to x86.



    Apple bought a company that has no history with x86. I think it unlikely they will switch their embedded products to x86 anytime soon.
  • Reply 95 of 105
    tenobelltenobell Posts: 7,014member
    Quote:

    I'm not a pessimist. I'm just saying that I don't agree with what some are saying here as to Apple's interest. I'm not denying that Apple has a purpose for this, and it may be a good one. But I think their purpose is different from what is often being said it is, here.



    The head of Apple's iPod division pushed for the acquisition and we have a quote from Steve Jobs saying Apple bought PA Semi because of its processor designs to be used for iPods and iPhones. I see no reason to believe that is not the immediate reason for the purchase.



    But of course PA Semi will provide chip designs for other products.
  • Reply 96 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by TenoBell View Post


    There isn't much Apple can do with in house GPU technology to improve its products or their sales. This is an entirely different story for embedded CPU's.



    That's not true. At the time, Apple could have made a difference.





    Quote:

    If Apple had bought a company that specialized in designing embedded x86 chips. Then I would agree Apple will likely move all of its products to x86.



    Apple bought a company that has no history with x86. I think it unlikely they will switch their embedded products to x86 anytime soon.



    [/quote]



    We really don't know what knowledge is there for x86. People are saying, without knowing, that PA and Apple have been working on an ARM chip for three years, or a PPC chip. It goes back and forth depending on the hour.



    It's just as likely they have gotten expertise in x86 during that time.



    Only time will tell, and it's kind of pointless to continue this, as I'm sure you'll agree, as neither of us know anything about it.
  • Reply 97 of 105
    melgrossmelgross Posts: 31,471member
    Quote:
    Originally Posted by TenoBell View Post


    The head of Apple's iPod division pushed for the acquisition and we have a quote from Steve Jobs saying Apple bought PA Semi because of its processor designs to be used for iPods and iPhones. I see no reason to believe that is not the immediate reason for the purchase.



    But of course PA Semi will provide chip designs for other products.



    We know they won't be providing any of their chips for the iPod, because they consume ten times as much power, and they will be discontinued, as PA has already announced.



    Apple won't be designing any new ARM's in the short term, because this company doesn't do any work with ARM's, only their own PPC designs.
  • Reply 98 of 105
    SpamSandwichSpamSandwich Posts: 30,837member
    Quote:
    Originally Posted by melgross View Post


    We know they won't be providing any of their chips for the iPod, because they consume ten times as much power, and they will be discontinued, as PA has already announced.



    Apple won't be designing any new ARM's in the short term, because this company doesn't do any work with ARM's, only their own PPC designs.



    Give us your best shot as to why this acquisition makes the most sense to Apple at this time.
  • Reply 99 of 105
    sc_marktsc_markt Posts: 1,393member
    Quote:
    Originally Posted by melgross View Post


    LCoS failed with Sony as well. Intel was just smarter about getting off earlier. Wasn't Intel's fault.



    One thing I believe, when Jobs said that they were "through" with PPC, he meant it.



    I just don't see Apple taking the huge expense, and risk, associated with doing a processor chip. The costs are almost impossible when you have no history. And PA's history isn't enough. Despite the ARM knowledge that is there for a couple of people, it's fairly old knowledge now. They would have to get up to speed, and the rest of the staff would have to learn even more.



    And for what? I just don't see it.



    Apple could easily sink another $500 million into doing a new ARM version just for themselves, with no guarantee it would be better than whatever Intel will have two years from now when Apple's chip would come on line.



    I just don't see it.



    I remember reading a lot of similar responses to the idea that Apple was going to switch to Intel.



    I don't see Apple going back to PPC either but I can't find any other reason that explains this purchase.
  • Reply 100 of 105
    mdriftmeyermdriftmeyer Posts: 7,229member
    Quote:
    Originally Posted by melgross View Post


    We know they won't be providing any of their chips for the iPod, because they consume ten times as much power, and they will be discontinued, as PA has already announced.



    Apple won't be designing any new ARM's in the short term, because this company doesn't do any work with ARM's, only their own PPC designs.



    I'll make this clear: They aren't using the Intel Atom platform.



    Being in contact with folks who know the fabs know they aren't using the Atom platform.
Sign In or Register to comment.