TSMC will continue to be Apple's sole chip producer for the 2019 'A13' processor

Posted:
in iPhone edited October 2018
TSMC will continue to be the only supplier of Apple's A-series chips in 2019, according to a report, with the iPhone component supplier believed to secure orders for next year's "A13" chip following its success with this year's A12 production.




The exclusive supplier of A-series chips since 2016, murmurs within the supply chain expect Taiwan Semiconductor Manufacturing Company (TSMC) to continue being the only firm producing the Apple-designed chips.

According to Digitimes, TSMC will fulfill all "A13" orders, and will likely help the company increase its market share in chip production. After acquiring a market share of 56 percent in the "global pure-play foundry market" in the first half of 2018, in part due to the A12, the Apple orders are thought by report sources to help propel the share to 60 percent next year.

TSMC continues to be a major supplier despite alleged efforts by Samsung to become one of Apple's A-series suppliers once again. Samsung was once the exclusive A-series manufacturer, but as competition and legal battles between it and Apple intensified, orders migrated over to TSMC, who now enjoys the monopoly of supply.

In June, Samsung was claimed to be in full development of an integrated fan-out (InFO) packaging technology and had mastered 7-nanometer extreme ultraviolet lithography (EUV), which would compete against TSMC's own 7-nanometer process. TSMC's InFO technology is said to make its process more competitive than its counterparts, with the company also tipped to announce the first commercially-available 7-nanometer EUV process, ahead of Samsung.

The relative lack of processor producers who work at the 7-nanometer level means there is little competition for TSMC in the field, aside from Samsung. The shift down to a 7-nanometer process is a difficult and costly undertaking for chipmakers, with the majority steering clear of it for the moment.

Both Qualcomm and MediaTek have postponed their own 7-nanometer process launches until 2019, according to reports. Globalfoundries has put its 7-nanometer FinFET technology development on hold, while UMC has shifted its focus to "mature" and speciality process nodes.

Meanwhile, TSMC's mastery of 7-nanometer production is expected to garner orders to produce chips under contract for AMD, Huawei, MediaTek, Nvidia, and Qualcomm.
cincymacclaire1

Comments

  • Reply 1 of 17
    Guess we should congratulate TSMC for a job well done.
    olsappleheadTomEclaire1
  • Reply 2 of 17
    GeorgeBMacGeorgeBMac Posts: 11,421member
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    edited October 2018 appleheadclaire1
  • Reply 3 of 17
    wood1208wood1208 Posts: 2,905member
    TSMC and Apple together have done great job pushing chip boundaries. A13 will built on the success of A12 with performance improvements. With A13, next year 2019's iPhone XS/MAX possibly can have triple camera,5G(Intel XMM 8000 modem),6th gen WiFi(802.11ax),etc and XR will inherit 2018's dual camera,Gig LTE from XS. Than in 2020, It's time for A14 to graduate from iPhone and move into Macbook.
    edited October 2018 appleheadclaire1
  • Reply 4 of 17
    aplnubaplnub Posts: 2,605member
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    ronnGeorgeBMac
  • Reply 5 of 17
    thttht Posts: 5,420member
    aplnub said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    Doing it this way, both processors in the machine for awhile, implies the marginal cost of a CPU competitive ARM is zero. Likely not, and it will mean higher prices per machine.

    The biggest impediment to a CPU architecture switch for a PC is software compatibility. That really can’t be done without a forced migration plus offering developers a larger market to sell into to make the development costs worth it. Windows/ARM failed for this reason and will continue to fail for this reason.

    If it is done, it will be like the PPC to x86 switch, or the 68k to PPC switch, they’ll offer separate machines, one ARM, one Intel, and phase out the Intel ones over a 2, 3 maybe 4 years.

    The writing is also on the wall. Or it is a big trial balloon. Apple wants developers to move their apps to Marzipan. If they do a good job, an x86 to ARM switch will be seemless. So, 2020 if Marzipan is the number one thing coming out of WWDC19. Maybe Fall 2019 for a first PC style machine with an Apple ARM. It will depend on how well Marzipan can accommodate complex apps like FCPX or Xcode. If you see Apple announce Xcode using Marzipan, the writing is burning on the wall telling developers than x86 will be phased out.

    The other option, which isn’t going to happen, is to force developers to architecture independent software like HTML/JS, or a Java style clone. Apple will have hard enough time to move developers to Swift let alone a lower performance language that will make certain apps much less performant.
  • Reply 6 of 17
    melgrossmelgross Posts: 33,510member
    aplnub said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    X86 isn’t a coprocessor today. The chip Apple uses is just for peripheral uses. Let’s not go too far in our enthusiasm here.
    GeorgeBMac
  • Reply 7 of 17
    melgrossmelgross Posts: 33,510member
    I am impressed by what Apple is doing. I highly recommend anandtech’s review, and deep dive of Apple’s new phones. It seems that Apple is selling itself short with the performance of the new A12.
  • Reply 8 of 17
    thttht Posts: 5,420member
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   
    As the adage goes, developers and users will find a way to use all the computational power in a machine. As long as a smartphone has general purpose computing functions, all computing power will be used up.

    Computational photography is a new buzzword feature. This type of thing will continually suck up all the computational performance that is available to it. Web page apps will continue to soak up all the CPU power possible. There can never be enough CPU and bandwidth available for ads which is 99% of what most web pages are today. Apple as created an AR file format. That can eat any devices lunch. In the future, VR functionality will require even more. You can just keep on going.

    For PC functionality, the only difference will be larger thermal envelopes will allow for more computing power.
    GeorgeBMacclaire1
  • Reply 9 of 17
    GeorgeBMacGeorgeBMac Posts: 11,421member
    tht said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   
    As the adage goes, developers and users will find a way to use all the computational power in a machine. As long as a smartphone has general purpose computing functions, all computing power will be used up.

    Computational photography is a new buzzword feature. This type of thing will continually suck up all the computational performance that is available to it. Web page apps will continue to soak up all the CPU power possible. There can never be enough CPU and bandwidth available for ads which is 99% of what most web pages are today. Apple as created an AR file format. That can eat any devices lunch. In the future, VR functionality will require even more. You can just keep on going.

    For PC functionality, the only difference will be larger thermal envelopes will allow for more computing power.
    ALL true!

    But another facet that could influence the impact of an A13 in an iPhone is the concurrent inclusion of 5G communications.   The combination of the two (CPU power and communication enhancements) have driven most of all of the major advances in computing since there has been computing.

    (The other factor has been storage capacity and cost -- particularly after it transitioned to "real time" (disk) storage from mag tape.)
  • Reply 10 of 17
    Good for TSMC! I’m happy to see Apple avoid giving any money to Samsung. I won’t buy any of their products anymore. (Which makes buying a cheap TV tough.)

    As far as the Switch to ARM, that’ll go MacBook (possibly with a special version with pencil support) then Mac mini and iMac models. Then the next year the MacBook Pro, iMac Pro and Mac Pro will move over. 
    claire1
  • Reply 11 of 17
    wizard69wizard69 Posts: 13,377member
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I kinda doubt Apple will call the “A” series.    What I’m expecting is a “B” series better optimized for the laptop and desktop markets.   The B’s would have more high performance cores and maybe as little as one low power core.  (The low power core to enable always on at less than a watt).   The B’s would also have other desktop and laptop oriented features to set it apart from A series. This isn’t significantly different from what Intel and AMD do with their chips.   In Apples case with ARM they have the potential to put a lot of cores in such a chip.  Also they could implement the ARM/Fujitsu  vector instructions that would blow away X86 processors.  
    claire1
  • Reply 12 of 17
    wizard69wizard69 Posts: 13,377member
    tht said:
    aplnub said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    Doing it this way, both processors in the machine for awhile, implies the marginal cost of a CPU competitive ARM is zero. Likely not, and it will mean higher prices per machine.

    The biggest impediment to a CPU architecture switch for a PC is software compatibility. That really can’t be done without a forced migration plus offering developers a larger market to sell into to make the development costs worth it. Windows/ARM failed for this reason and will continue to fail for this reason.

    If it is done, it will be like the PPC to x86 switch, or the 68k to PPC switch, they’ll offer separate machines, one ARM, one Intel, and phase out the Intel ones over a 2, 3 maybe 4 years.

    The writing is also on the wall. Or it is a big trial balloon. Apple wants developers to move their apps to Marzipan. If they do a good job, an x86 to ARM switch will be seemless. So, 2020 if Marzipan is the number one thing coming out of WWDC19. Maybe Fall 2019 for a first PC style machine with an Apple ARM. It will depend on how well Marzipan can accommodate complex apps like FCPX or Xcode. If you see Apple announce Xcode using Marzipan, the writing is burning on the wall telling developers than x86 will be phased out.

    The other option, which isn’t going to happen, is to force developers to architecture independent software like HTML/JS, or a Java style clone. Apple will have hard enough time to move developers to Swift let alone a lower performance language that will make certain apps much less performant.
    I really don’t buy into much of anything above!    

    For one Apple will simply recompile XCode to run on ARM.   They will expect developers to do the same.   Beyond that they have already built into the app stores the systems required to get the right machine code to the platform it is installing on.  

    As for MS supposed failure with ARM that is more an issue with MS and their ecosystem.   The Linux world has had zero problems with Linux on ARM when it comes to ecosystem acceptance.   Considering the average developer in Apples ecosystem is smarter than the average developer in Micro Soft land I don’t see Apple having any issues at all with software support on ARM.     

    Beyond that you need to consider how the vast majority of Apples customers use their machines.  The overwhelming majority of the apps would be native on day one.  
    claire1
  • Reply 13 of 17
    wizard69wizard69 Posts: 13,377member
    melgross said:
    aplnub said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    X86 isn’t a coprocessor today. The chip Apple uses is just for peripheral uses. Let’s not go too far in our enthusiasm here.
    I think that depends upon you perspective.  As I understand it Apples chip controls the boot process, some would see that as the X86 being a slave to the Apple chip. To make it a coprocessor all you need is to have Mac OS running on the Apple chip. 
  • Reply 14 of 17
    claire1claire1 Posts: 510unconfirmed, member
    Good. This just means we won't hear obnoxious knockoff iPhone fans say "Samsung makes the chips in the iPhone!!! LOL"

    Most stupid logic ever.

    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    Not to sound mean but use some imagination. This is a pet peeve of mine. I remember people arguing that graphics wouldn't go beyond the Xbox 360 capabilities...
    I can think of a few features that will stump current iPhone.

    How about shooting 3D VR content for interactive entertainment?
    Console-level games?
    Music creation?
    3D Animation for movies?

    Facebook browsing was near flawless on the first iPhone but time has passed and Facebook/other services will continue to evolve and add more features. I'm sure people in the 90s couldn't comprehend why computers would need more power. Wasn't it Bill Gates who said we wouldn't need a gig of RAM?
  • Reply 15 of 17
    melgrossmelgross Posts: 33,510member
    wizard69 said:
    tht said:
    aplnub said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    Doing it this way, both processors in the machine for awhile, implies the marginal cost of a CPU competitive ARM is zero. Likely not, and it will mean higher prices per machine.

    The biggest impediment to a CPU architecture switch for a PC is software compatibility. That really can’t be done without a forced migration plus offering developers a larger market to sell into to make the development costs worth it. Windows/ARM failed for this reason and will continue to fail for this reason.

    If it is done, it will be like the PPC to x86 switch, or the 68k to PPC switch, they’ll offer separate machines, one ARM, one Intel, and phase out the Intel ones over a 2, 3 maybe 4 years.

    The writing is also on the wall. Or it is a big trial balloon. Apple wants developers to move their apps to Marzipan. If they do a good job, an x86 to ARM switch will be seemless. So, 2020 if Marzipan is the number one thing coming out of WWDC19. Maybe Fall 2019 for a first PC style machine with an Apple ARM. It will depend on how well Marzipan can accommodate complex apps like FCPX or Xcode. If you see Apple announce Xcode using Marzipan, the writing is burning on the wall telling developers than x86 will be phased out.

    The other option, which isn’t going to happen, is to force developers to architecture independent software like HTML/JS, or a Java style clone. Apple will have hard enough time to move developers to Swift let alone a lower performance language that will make certain apps much less performant.
    I really don’t buy into much of anything above!    

    For one Apple will simply recompile XCode to run on ARM.   They will expect developers to do the same.   Beyond that they have already built into the app stores the systems required to get the right machine code to the platform it is installing on.  

    As for MS supposed failure with ARM that is more an issue with MS and their ecosystem.   The Linux world has had zero problems with Linux on ARM when it comes to ecosystem acceptance.   Considering the average developer in Apples ecosystem is smarter than the average developer in Micro Soft land I don’t see Apple having any issues at all with software support on ARM.     

    Beyond that you need to consider how the vast majority of Apples customers use their machines.  The overwhelming majority of the apps would be native on day one.  
    There’s no such thing as a simple recompile. only if the software is really simple itself does it work. Otherwise, there remains months of work. It could be much harder in this case, as ARM is missing a lot of the instructions of x86, for which there are no counterparts in the ARM architecture. So this is oversimplification.

    in reality, Apple will have to deal with this problem, and yes, it is a problem. Just moving from Carbon to Cocoa cost Adobe, Microsoft and a number of other large developers, including Apple itself, quite a bit of time.

    i still think the best solution is to add a number of x86 instructions to their ARM SoC for this purpose. Individual instructions are not patented, or copyrighted. They’ve been used by numerous other processor manufacturers over the years. I want to make it clear, that those instructions aren’t actually x86 instructions, as they been used elsewhere. They are just instructions that Intel puts in the chips.
    edited October 2018
  • Reply 16 of 17
    melgrossmelgross Posts: 33,510member

    wizard69 said:
    melgross said:
    aplnub said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    X86 isn’t a coprocessor today. The chip Apple uses is just for peripheral uses. Let’s not go too far in our enthusiasm here.
    I think that depends upon you perspective.  As I understand it Apples chip controls the boot process, some would see that as the X86 being a slave to the Apple chip. To make it a coprocessor all you need is to have Mac OS running on the Apple chip. 
    You’ll have to explain how that would work. What would be the purpose of that?
    edited October 2018
  • Reply 17 of 17
    MacProMacPro Posts: 19,718member
    wizard69 said:
    melgross said:
    aplnub said:
    The interesting part of these chips is:   Since their primary use is in smart phones, and smart phone functionality is being maxed out (How fast can one browse FaceBook?), is there even a need for this level of performance in their traditional market of smart phones?

    My bet is that the power these chips are capable of will be opening up the tablet and laptop markets to the A Series processors...   That's bigger, much bigger, that Apple switching from Motorolla to PowerPC to Intel.   It will be like the marriage of two giants -- so that all that will be left separating them will be the software.

    (That said, I still expect the A series processors to be introduced in the MacBook line rather than the Pro line.  The Intel processors and high end GPUs still have an edge in the ultra high performance market.)
    I expect both Intel and A series processors to be included into a Mac at some point in the near future. Intel is just a coprocessor even today, and Apple will slowly diminish the perception of having an Intel processor. Going solo with A series only may still be a ways off.
    X86 isn’t a coprocessor today. The chip Apple uses is just for peripheral uses. Let’s not go too far in our enthusiasm here.
    I think that depends upon you perspective.  As I understand it Apples chip controls the boot process, some would see that as the X86 being a slave to the Apple chip. To make it a coprocessor all you need is to have Mac OS running on the Apple chip. 
    Maybe at some point, an Intel CPU might just be a BTO option for those that need Boot Camp.
Sign In or Register to comment.