Blog Post

Apple switch from Intel chips? The real question may be when, not if

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Once again there’s a rumor about Apple exploring plans to dump Intel(s intc) chips on its Macbook line of computers. The rumor is a favorite, but the question is probably less a matter of whether or not Apple (s aapl) is weighing a switch and more about when such a switch would happen. From a chip point of view, the stars are lining up, but they aren’t in formation just yet.

The combination of the overall change in computing, from something that occurs at a desk to something that happens on a variety of devices throughout the day, and higher performance chips based on the ARM(s armh) architecture planned for next year are probably the spur for the latest rumors. The change in computing has been happening for some time, but only last week did ARM announce its next generation processor cores, the A-53 and the A-57. The big focus of those cores was that they can support 64-bit computing — a necessity for servers and the personal computer markets today.

Better chips mean a better experience

In a conversation last week with Nole Hurley, VP of marketing and strategy for ARM’s processor division, we briefly discussed the likelihood of one of ARM’s new A-50 cores having the chops to make it into laptops and other devices dominated by CPUs. The 64-bit compatibility associated with the next generation ARMv8 architecture means that the cores will be able to address more memory. That gives devices the ability to support having scores of open tabs on their browsers while running a dozen different programs or apps.

Until these cores were announced, ARM cores have supported 32-bit processing. That’s fine for mobile devices where ARM has dominated, but when it comes to content creation, 64-bit compatibility is crucial, as you can see from my colleague Kevin Tofel’s review of a Chromebook running on an ARM chip versus one running an Intel chip. Thus, with these new cores, ARM has a product that could possibly compete in the laptop and personal computing market. Those cores will be out next year with devices running that IP expected in 2014.

If Apple is contemplating this switch it is because it realizes that now that computing has gone mobile, the x86 architecture, which was optimized for performance at all costs, has been superseded by the ARM instruction set that optimizes energy efficiency at all costs. The difference now is that ARM has also been boosting performance while optimizing for efficiency and has reached a level of performance parity that supports today’s application needs. Intel’s focus on efficiency in x86 is kind of like retrofitting a Ferrari to be more fuel-efficient while ARM has been building a Tesla.

Apple hasn’t stood still however waiting for better ARM cores. It has built a better core using its ARM architecture license to tweak the current generation ARMv7 instruction set to increase the performance of the A6 processor in the iPhone 5. (An architecture license lets Apple tweak the core designs that ARM builds.) As this deep dive from Anand Tech illustrates, Apple has already made a processor for phones that could rival a traditional CPU.

Apple hasn’t forgotten about the need for software

And Apple may have already started down the path of optimizing its OS X Mac operating system for the ARM-based architecture, one major step it needs to take before any transition off Intel’s x86 platform. Back in 2007 Apple said the iPhone runs OS X and Apple later clarified that it runs on the same kernel software. Thus, must of the optimization needed to fully retrofit OS X for the ARM instruction set may already be done. That only leaves the popular software running on the MacBook left for a retrofit, and because Apple already has adapted popular programs of its own such as iMovie and iPhoto for iOS, that work may also be done.

The biggest holdout is probably Microsoft Office, and one would assume that Microsoft(s msft), once it got word of Apple’s plans it would act to preserve that software on the popular Macbook platform. However, Microsoft has been dragging its feet with ARM. For example it supports Windows RT and has a mobile platform that runs on ARM chips, but Windows 8 still doesn’t.

So, as the rumor mill works over the possibility of an Apple switch from Intel to ARM-based chips in the Macbook, it’s worth noting that the hardware is almost in place and the software is probably mostly there. The biggest unknown in this plan is how Intel might react to the threat of an Apple defection.

32 Responses to “Apple switch from Intel chips? The real question may be when, not if”

  1. Haha. Just wait, I’ve got a crazy idea that Apple might buy AMD is get control of their processors rather than build it all themselves from scratch.

    Not only would they get Intel compatible parts, but they’d also get the 64-bit ARM tech that AMD is working on as well as ATI and all the graphics coprocessor technology (that they could use for both mobile and PC products). With AMD’s price currently in the toilet, This could be a great strategic purchase for Apple.

    About Apple’s relationship with Intel:

    1) Intel and Microsoft are still close allies. Microsoft has reemerged as a competitor that Apple dislikes. Their Oct 23rd product announcements only talked about 1 competitor, and it was Microsoft.

    2) Apple ships approximately 13% of the PCs manufactured annually. While that doesn’t dominate the OS usage, it DOES make Apple the biggest individual PC manufacturer. That means they’re paying billions to Intel quarterly for processors. This is probably their biggest external expense by far.

    3) Apple doesn’t like ANY parts to be single source. They are SCREWED if Intel starts to give them grief. And stories about Apple dumping Intel (for any reason) will upset Intel and have them rattling sabers at Apple in a heartbeat.

    4) Haha. and of course, Apple doesn’t seem to like the idea of others getting any part of the money for their products. ;-)

    5) Apple completely rejects all of Intel’s MDF dollars and branding programs. The new Retina Mac 13 SHOULD be able to so WiDi wireless display since it’s using all the correct Intel pieces, but Apple doesn’t include the software, nor will they add the WiDi logo to their products. Similarly, Both Retina Macs qualify as ultrabooks, but Apple won’t participate in that branding programs. So they leave millions of dollars of MDF funds on the table.

    And I haven’t even mentioned the new Trinity (FM2) stuff from AMD that combines a full ATI graphics controller AND AMD processor on the same chip for higher performance and lower power usage, this is exactly the kind of stuff Apple loves. Plus the Trinity processors are a LOT less expensive than Intel ones, so it could help Apple get their price down without sacrificing their high margins.

    Overall, I think it would be crazy NOT to consider that Apple has it’s eye on AMD! ;-)

    • On 2)
      Simple layman math – How is ( Buying cost of AMD + Quarterly Operating Cost + Manufacturing cost ) stack up to ( Quarterly Purchase cost of Processors) ?

      Its true that you have more control and eliminate dependencies if you build everything yourself , but what is the cost advantage? Companies like Intel and AMD doesn’t simply manufacture chips based on somebody else’s design(could have exceptions). But they do a quite lot of R&D and new ways of designing and manufacturing. And they specialize on those. Its not like buying “Sparrow” or companies involved in Maps.

      Why are people making predictions based on the Cash Reserve of Apple and Market Value of a competitor or a Company? Have seen quite a lot of rumours based on this.

    • Chris Meredith

      #5 coincides with the point i was making earlier, however just to be the anal retentive person i am , Apple cannot use WiDi, because it uses the Broadcom Wifi/BT chipset. I was going to put a Win7 VM on my Mac and load all the drivers to see if I could kludge WiDi on my MBP, however I found out that Apple doesnt use the same interconnect as most of the market for it’s Wifi Card.

      I have been able to kludge it on a Hackintosh :-)

      your points are well taken, I think that if Apple took the MDF $$$ it would make them feel like they are diluting the Apple Brand and that Mr Jobs didnt invent everything. Tim Cook did say he was going to buy things with the $1000 Billion dollars Apple has in the bank

  2. Ian Abbott

    The main problem will be for developers… one of the reasons that Apple laptop sales went through the roof after the Interl switch over was the easy ability to develop for four platforms (OS X, iOS, Windows and Linux) on one machine. Whether it was UNIX system tools programming, cross-platform applications or web development, the Macbook line could handle it. Switching to ARM effectively kills many people’s usage of the computers. Sure, they could use some translation system a la Rosetta for OS X apps, but when it comes to other Intel-compatible platforms? Nope. We’d lose the quick virtual machines we’ve come to know and love and go back down to the boggy performance of emulated CPUs such as Virtual PC back in the PPC days. This would be a major mistake. Apple may find it simpler to unify their lines, but I don’t think they understand *why* so many people buy into their computer lines. It’s not just for OS X. It’s for a light, fast and capable Intel platform.

  3. Jeffrey Chan

    Pundit talking out of their asses again.

    People spewing these nonsense are applying the iOS business model to the Mac model. But that’s just plain stupid.

    1. The pro market is the butter and butter of the Mac ecosystem. Ableton Live, Final Cut Pro, Adobe CS. Switching CPU architecture potentially triggers a mass migration of Pro users. Especially the web developers. Most servers are Intel, if Mac switch CPU architecture, they can’t compiled code on their laptop anymore. VM will then have to do full x86 emulation and it be dog slow. You can also kiss the whole enterprise adaption goodbye. Few business executives will buy a Mac anymore once you take away BootCamp and they can’t run Windows as an option anymore.

    2. iOS SKUs uses all but 2 ARM chip SKUs. On the Mac, a chip maker will have to provide CPUs of various different clock rates, performance and power consumption levels. Manufacturing these different CPU configs divide production volume for each SKU and drastically increases manufacturing cost for these chips.

    3. And that was the WHOLE DAMN POINT of switching to intel in the first place: to exploit Intel’s huge volume in chip shipments so to keep the cost of the CPU part low. Even when Apple is only ordering 100,000 units of 2.7GHz Quad-core Core i7, Intel is also making millions of them for other OEMs, that’s what keep prices down for Apple. Making 100,000 units of A50 for Apple pretty much require Apple contracting their own fab just to make them, it’ll cost at least double the Intel counterparts.

    4. Apple’s been down that road before. The PowerPC chips Apple used were largely designed by Apple (IBM has their own PPC design for their workstations), contract fabs to make PPC that have no other markets other that inside the Macs. Not only manufacturing cost went up steadily, Intel time and again were able to refine smaller and smaller feature size (nm) before PPC did. At the end of PPC era, they were about 2-3 generations behind in nm, and still cost 50-100% more.

    5. iOS devices are consumer devices sold by the millions in the first week. A MacBook or iMac of a particular SKU might reach a million in one year. The economy of scale is vasty different.

    People who retell these BS don’t have the first clue about the cost-volume relationship of CPU manufacturing. They just blindly assumes that if Apple manufactures their own chips for the Mac, it’ll cost less for them.

    It’s the opposite. Apple had already done it BEFORE. that’s why they switched. If you read the original switch announcement, Apple specifically says that, the reason for the switch is exactly because Apple can buy CPUs that Intel is already shipping in huge volume so they are keep cost down.

    None of these pundits had a Mac before the switch. They have no damn idea how painful that is to developers and users. They think it’s just another “business decisions”. Guess that’s why they’re pundits rather than running companies.

    • Jeffrey Chan

      BTW I apologize my post wasn’t in the best tone. I’m just tired of year after year people dig this back out, and talk about a CPU architecture switch like a nothing but a business decision, when such move doesn’t even make any business sense in the first place. It just boggles the mind

      Really didn’t mean to offend anyone.

      If you remember how Apple had to pay 2x the price of an Intel chip for a G5 part @ 150 sqmm die fabbed @ 130nm when Intel/AMD was already on 90nm @ <100 sqmm, when said G5 that cost 2x as much can't even deliever 60% Intel/AMD performance and require liquid cooling, you'll be upset to when you hear people saying switching is anywhere remotely close to a sensible business decision.

      Intel chips ship hundreds of millios a year, they'll ALWAYS have a much tighter revision cycle and cost MUCH less. The fancy new 10nm fab will always pick up the 100m order first before they even look at your tiny 1m order.

    • Refreshing to hear some sense in the comments at last. However nice the ARM technology may be, there are real issues with switching away from Intel – nobody has yet mentioned the fact that x86 compatibility means the ability to boot Windows (directly or as a virtual machine), which is a huge factor in Windows users switching to Mac. Just having the ability to use Windows occasionally is a big win that Apple would have to throw away to move to ARM.

    • Chris Meredith

      interesting article, Apple probably has OSX running on ARM, and they quite possibly are contemplating the Switch. The way apple works means they should think about it, it would allow them to control the experience that the customer has on the platform.

      there are a number of things to consider, pros and cons for this scenario

      Jeffrey Chan pointed out that the Pro Market has supported Mac during the dark days , and I believe that to be true once upon a time, yet i wonder if it is still true? I wonder as Apple has not refreshed (significantly) the Mac Pro Line in ~18 months. The Last release of Final Cut Pro was ghastly. So does Apple support the Pro Market?
      Running a VM to enable X86 compatibility may be an issue, or it may not, I read recently a Russian company is working on on-the-fly x86 compatibility for ARM ala Transmeta, so this may become a feature that if it doesn’t take too many cycles may remove the need for a VM, just a licensing agreement

      Enterprise adoption is another issue, this seems to be being driven by the BYOD crowd ( I may be wrong, however GigaOm has many articles a month on this being the case. Windows RT supports ARM, Office 2013 Supports WinRT on ARM, eventually this will be a non-issue if the take up rate on WinRT ARM gets high enough for MS to throw their weight behind it (remember MS is now a HW company)

      Intel and Apple (based on news reports) have a somewhat stormy relationship, however they have both benefited from this relationship. Intel has given exclusivity to Apple (I.e. Thunderbolt, Ultra LV Processors for MBA,) and Intel has certainly been helped to gain market share by the relationship. There are a number of widgets that Apple doesn’t use that Intel provides to other companies, Quick Assist/Cave Creek, would make trans-coding much faster for video pro’s and other more valuable features too numerous to mention.The Full Intel 4000 Embedded GPU, this GPU by default supports 4K video, 8 simultaneous Displays and many other neat features, and future versions of Intel Embedded graphics will eventually catch up to the Discrete graphics and these are cores in the CPU , so they will take advantage of dies shrinks same as the CPU.

      Atoms SOC’s may eventually be powerful enough to power a Mac Laptop, but why would Apple do that when Haswell is coming out in late 2014/2015? Double Ivy bridge performance in 17 watts or Ivy bridge performance in less than 17watts. Will ARM be able to match this in 5 years?
      ARM is currently the leader in power and long battery life, will these be the case when all of the 64 bit extensions, caches, pipelines out of order execution and other features are not currently present in the ARM cores? Where will Intel be with Power consumption in 5 years? Intel’s Main problem is they have a leaky architecture, one hopes they will have made significant strides in reducing this by 2017!

      Does Apple have the expertise to add all these features to their SOC, probably, if not now they will, if not they may cross-license AMD. I doubt Intel will ever become a foundry for Apple, Not in there best interest, nor would it be cost effective for Intel. It would cost Intel more money to Switch the production line back and forth from X86-ARM and back again then they would probably make in the deal.

      The only people who care about all of this mental masturbation are the pundits, and the executive/employees of the associated companies. I switched to a Mac in Mid-2001, Was it because of the HW capabilities? The coolness of the brand, the exorbitant costs of the system?
      it was none of the above, it was the OS combined with features of the SW , that worked better than the MS Windows /SW that drove me to change. It was features that saved my time and allowed me to be more efficient, the first couple of tries for OSX , were problematic, but now with ML it just works. what we as people who are knowledgeable of the inner workings of the computer sector seem to forget, is that most people who use computer/tablets/phones et al , is that they want them to have certain features and those certain features to Just work. if it works and is slow they will buy something faster.

      Apple will most likely (in the future) switch to using ARM cores, Intel will eventually design processors that are energy efficient and powerful. I just hope that OSX just works, it fulfills the promise Linus made in the 1990’s, that i will be able to run Open Source SW, Commercial SW , have a nice Gui and still have a powerful CLI . So as long as they don’t Ubunt-ize or MS Windows OSX i will be happy and both Intel and Apple can continue their journey to take over the world

  4. This interest in switching from Intel to ARM just proves that Apple is no longer a computer company (as they dropped that part of their name a few years ago) and are just a consumer gadget company.

    For Apple it is not about putting out the fastest computer or most upgradeable computer as they lost that title years ago when they started to solder the ram onto every computers motherboard so you can’t upgrade them after you buy them from Apple.

    iOS or bust for everything and even then iOS is the oldest, least multi-tasking of any mobile/tablet OS out there but who cares as you have billions of dollars of marketing behind things as well as the always positive press to help sell the gadgets.

  5. Stacey – Your article captures well a what-if scenario when Apple takes this route.

    Rumours say Apple may take this route to technically advance the current Mac Desktops/Laptops. Apple representatives them selves have been hinting or projecting a Post-PC Era. Why would Apple invest heavily, as it is required, to move towards that Post-PC Era. 2017 – what’s projected, by then no one can predict whether laptops and Desktops will exist , even if they do in what form.

    Why not collaborate with Intel or other Chip manufactures and let them manufacture chips optimized for Apple?

    Will Apple also start designing their own hard disks to have a minimum form factor for their laptops??. It would be foolish to design each and every component that goes to Apple Products. What happens if the heavy investment Apple make, goes useless due to technology advancements coming in the near future. And this will make the entire system complicated. Currently Apple needs to concentrate on the Overall Design of their products and the Software that runs on it. Would they want to move away and micro design each and everything that goes into their product?

    • Jeffrey Chan

      Intel has been making CPUs “optimized” for Apple since the first MacBook Air. It was an ultra low voltage SKU, lower than any other Intel CPUs at the time and was NEVER on Intel’s roadmap until Apple approached Intel about it.

      • That was news to me. Very well agree to your comments on cost factor. Its better to leave the component manufacture to others as the Cost factor they can achieve cannot be met by going by the demand Apple has.

  6. Some fundamentally flawed arguments made in there:

    *Yes ARM will get 64bit soon meaning more memory, this does not translate directly to better performance as was always demonstrated by the PowerPC era of processors.
    *The ARM chromebook didn’t perform as well as the Intel one because it’s significantly lower power, not because it’s 32bit
    *ARM processors are usualyl significantly lower wattage than Intel which means better battery life, but they’re also significantly less powerful than Intel at the same clock speeds.
    In fact with current Ivy Bridge processors, Intel is challenging ARMs battery life domain while providing a lot more performance

    *While OS X may be able to switch over to ARM relatively easily, ditching x86 is a significant endeavor. It’s not just the OS that needs to transition over but every single app that a user uses.
    Unless Apple has a system like Rosetta in place, it means users won’t have anything to run except for iPad apps.

    *Office 2013 is already available on ARM
    *Windows 8 is Windows RT in some sense. It just locks out x86 desktop apps on ARM processors, but most if not all Windows Store apps will run equally on both architectures.

    • H. Murchison

      I believe the reference to 64-bit was about addressing more memory. Also 64-bit can deliver processing advantages as well but merely having a 64-bit processor means little if you don’t have 64-bit software to take advantage.

      The ARM Cortex A15 in the Chromebook is about $30. So we’re comparing a $30 SoC to an Intel chip that’s likely $200 or more.

      ARM’s approach is going to be scaling multiple processors. The lower wattage becomes important here.

      Intel may challenge ARM in battery life but that is of no consequence to Apple. They want low battery life but they also want to be able to tailor the Processor for the traits they need like Memory Bandwidth or GPU performance. Intel’s offerings are “off the shelf” you get the same product as all of your competitors.

      Moving to Intel was difficult but software design and tools has improved greatly. Moving off of Intel isn’t going to be as hard as the transition to it. Most major Apple Frameworks are already on ARM. AV Foundation, Core Audio/Video, GLKit, etc.

      OS X portability is far beyond what Microsoft has and that’s understandable considering the size of each companies respective install base.

  7. H. Murchison

    “Apple needs Intel more than Intel needs Apple”

    AAPL market cap = $549.44 billion
    INTC market cap = $118.18 billion

    Apple almost could buy Intel with their cash assets. The days of Apple needing Intel evaporated when the majority of revenue and profits came from ARM based devices.

    But forget that let’s talk technology. If you develop applications for OS X you are never writing to what is called “bare metal” so for all intents and purposes you are writing code that is platform agnostic to a certain point. But it gets better. Apple has moved away from the venerable open source compiler GCC to a more flexible solution called CLANG/LLVM. This compiler makes it easier to target your code to the proper architecture. In fact a universal binary can be delivered in Xcode that works on Intel or ARM hardware.

    Take home message is that moving from Intel to ARM would be a far easier transition than moving from PPC to Intel which had Endian issue and some “bare metal” issues.

    One the benefits of the ARM cores is power efficiency and core simplicity. ARM is moving to 64-bit but those aforementioned traits should still exist. Here’s the kicker though. ARM has a very fast interlink for the cores appropriately named CoreLink GIC-500. This means as a hardware vendor I can simply daisy-chain the cores together and have them share cache resources and more quickly. The future is scalable processing power. It’s happening in the datacenter and will filter down to the desktop faster than many can imagine.

    We know this design of scalable works. Sony uses their Synergistic Processing Elements(SPE) in the Playstation 3 and your basic GPU is based on many simple cores rather than a few complex cores a la Intel.

    The key moving forward is going to be seeing how iOS and OS X evolve especially towards this sort of parallel computing. Apple has open sourced Grand Central Dispatch but they’ll need even more robust technology to handle this sort of scalable ARM system in the next half decade if they wish to supplant Intel.

    Great article Stacey

    • Tihomir Katulić

      “Apple almost could buy Intel with their cash assets. The days of Apple needing Intel evaporated when the majority of revenue and profits came from ARM based devices.”

      This still holds true if you take into account the desktop/laptop market. Unless they switch to AMD or VIA.

      “Apple has open sourced Grand Central Dispatch but they’ll need even more robust technology to handle this sort of scalable ARM system in the next half decade if they wish to supplant Intel.”

      My point exactly. 5 years, and then we’ll see it – or we won’t. Cause Intel will be standing still, like Nokia after N95 in 2006., while Apple co-develops all that ARM goodness. Yeah Apple has a lot of cash. So has Intel – maybe they’ll go for ARM?

      Personally, I would like to have more choice on the desktop/laptop market. It would be great to have a competing platform, and ARM sure looks like the most promising competitor, but as things are at the moment, I don’t see myself retiring my x86 laptop let alone the workstation desktop I use. I like my Ipad and my Iphone as much as the next guy, but lets not get carried away.

    • Jeffrey Chan

      That’s exactly what upset me how people discuss architecture switch. What does market cap have anything to do with this? Shipment is everything in CPU manufacturing and pricing.

      Apple can go buy 10 Intels and 20 AMDs, but at the end of the day, what keep CPU part price down for Apple is the other 200 million PC shipments that uses the same Intel chips.

      Macdonalds make millions of burgers a day, that’s why they can sell it at 99 cents or whatever they are now. While your local restaurant makes a few dozens a day, that’s why they sell it for $8-10, even when they hire less staff and have half the floorspace of the McDonalds next door.

      • H. Murchison

        Market cap signifies the financial strength of a company to a certain degree. I don’t know of any company that is tops in their field that wants to be at the mercy of another provider.

        If you look you’ll see that the Apple and Intel relationship isn’t that chummy. Apple delivered the Macbook Air and Intel turned right around and initiated their Ultra Book specification with the MBA copycat devices.

        Apple cannot and will not take the next step until they are controlling their own destiny. ARM isn’t fully controlling it but it’s a larger step towards more autonomy.

  8. There’s no doubt in my mind Apple will do this. They might even do it as soon as 2014, once their 64 bit chips are ready, or they might wait a little longer until TSMC’s 14nm FinFET process is ready in 2015, so they can increase the CPU’s clock to higher frequency (2.5-3 Ghz) while still keeping very low power consumption – and no fans.

    They could certainly do it as early as 2014, though, and with ARM GPU’s is already easy enough to support 2560×1600 resolutions, and even 4k resolutions starting with next year. I could see them doing a $699-$799 Macbook Air with ARM chips.

    But speaking of GPU’s, once they are committed to switching the whole Mac line-up (over a period of a few years, of course) to ARM, they’ll want to design their own GPU’s too, and have complete control over that part of the SoC as well. Sourcing Imagination is not good enough, and now that Imagination also plans to compete with Apple and others in CPU design with the acquisition of MIPS, they’ll want to distance themselves from them, just like they’re doing with Samsung. They will still use Imagination in the first ARM Macs, though, so this probably won’t happen until 3-4 years from now, at least.

    For people who still think Intel will continue to make the most powerful laptop chips – they are really missing the point. It actually surprises me that the most skeptical people about this move are actually Apple fans. They should know better. Apple doesn’t care that much about total performance, as long as they think something will make up for the lack of performance.

    Take the recent 13″ Macbook Pro with retina display, for example. They didn’t put a more powerful GPU in it because they cared about heat, battery life, and thinness. So they used Intel graphics chip – which shows they cared about those qualities more than the raw GPU performance. But they did this anyway, despite a much lower performance, to have the retina display.

    People have also forgotten that the first Macbook Air was also a much underpowered laptop, but they did it anyway to claim they have the lightest and thinnest laptop on the planet. For Apple these kind of qualities are much more important than total horsepower. Because they know it’s these qualities that sell the products, ultimately.

    And I won’t even go into how much Apple would love to control their own chips in the Mac line as well. Steve Jobs might’ve wanted it for this reason alone. And make no mistake – this is something Steve Jobs wanted, too, not just Tim Cook. They hired these ARM CPU engineers to make their own CPU years ago, and I’m sure they’ve been considering putting them in Macs ever since.

    Apple WILL use ARM chips for its Macs. You can bet on it. And it will probably happen much earlier than everyone is expecting – 2014/2015, rather than 2017 like in the source article.

      • They could still buy Imagination, which would be pretty cheap for them. But Imagination might not want that, now that they want to design MIPS CPU’s, and Apple would most likely kill that division if they bought them. Imagination seems to have higher ambitions than just selling GPU’s to Apple.

        It also seems Imagination will want to sell their MIPS CPU’s to Android manufacturers, because Android supports MIPS, so Apple might not like that. If Apple can’t buy them, and MIPS will increasingly deal with their “rivals”, then Apple might want to create its own GPU’s. I think they want to do that anyway. The first step was the CPU, the next step will be the GPU.

        We haven’t heard any rumors that are related to this, though (no GPU company acquisitions or hiring of GPU engineers), so it may take a couple of years before they even start working on it, and then a couple more years before they’re ready to ship it.

        If Apple really wants to unify their chips in all their products and have full control over them, and adding exactly the features they need to them (like they did with A6 core instead of getting A15, or big.Little), then it seems logical they will want to have full control over their own GPU’s too, in the future, which is also why they won’t use ARM’s Mali, to answer your question.

      • H. Murchison

        @Stacey @Lucian an AMD acquisition brings ATI engineers into the fold. AMD is dirt cheap right now and former AMD engineers Jon Bruno and Jim Mergard are already at Apple.

    • Jeffrey Chan

      How much is TSMC’s 14nm FinFET process gonna cost to make 1 million ARM SKUs that only Apple will use and absolutely nobody else?

      Intel shipped 75 million Sandy Bridge alone, Apple sells about what? 15-20 million Macs a year? And that’s spread over 8-10 SKUs. That’s about 2m units max per CPU stepping a year.

      You better have a pretty damn compelling reason to not only switch your OS to a CPU no other desktop computer use and make your users do another painful transition, and to fab your own chip at such low volume.

  9. Tihomir Katulić

    Dear Stacey,

    While the premise of the article is interesting, the analysis lacks a couple of important factors that would surely influence Apple’s decision to ditch Intel chip architecture.

    Switching architecture at this moment means switching operating systems and software. This is always a traumatic move, one that costs a lot of money and is done only when there is no other way (i.e. when Apple ditched the crappy OS IX in favor of the BSD *nix derivative we now call OS X) or when switching means a big step forward (Apple ditching Power PC chips for better and cheaper Intel chips, as well as access to NVidia/ATI graphics).

    Now, if ARM chips were faster than Intel x86 processors, that would constitute a good reason to think about such drastic change that requires users and developers to switch to a new OS (be that it might that the new OS is basically its popular tablet/phone OS). Unfortunately, this is not the case. ARM chips are still far, far away from meaningful and competitive desktop usage, and the performance gap is not shrinking. Quad core 1.4 GHz ARM chips in mobile phones and tablets are still embarrassingly inferior to x86 compatibles from a decade ago.

    So, unless ARM/Apple processor design catches up with Intel, if not completely, then at least sufficiently in order for them to offer a desktop user experience that doesn’t lag like a Pentium III running windows 7 without enough RAM, there will be no architecture change in the foreseeable future. It would be a bad move for Apple, as I’m pretty sure they know too well.

    Far more likely, if Apple really wants independence from Intel, they might turn to AMD, who can supply them with comparable CPUs and GPUs, and who in light of its recent problems might be more open to cooperation and Apple design demands than Intel.

    However, if I were to make a prediction based on publicly available facts today – the Intel ditching rumors are just that – rumors. Apple needs Intel more than Intel needs Apple.

    • I am going to disagree with you on this. While there will always be some applications and customers for high-end Intel-based systems Apple is emphasizing portability, thinness and battery life. Once ARM cores are able to address more memory I can’t see Apple throwing aside the power advantage ARM offers for performance that most apps today don’t need.

      Plus with the architecture license and the purchase of Intrinsity, Apple can optimize the ARM cores for performance in ways that other chipmakers cannot. So it’s not even a straight ARM to Intel comparison for Apple.

      • Tihomir Katulić

        Fair enough, we can agree to disagree.

        Granted, Apple wants to rely only on in house developed technology, but currently I just don’t see it happening anytime soon, not on the laptop/desktop front. They could probably build a MB Air with ARM chip right now, but with the tech they got, it would be hideously outclassed by a 99$ Chinese netbook running on Atom. I’m sure someone would buy it though…

        So there’s only one way to settle this – let’s come back to this thread in a few years and see whether Apple ditched x86 or not :)

      • Jeffrey Chan

        You are basically suggesting that Apple will ditch the Pro market for this, it’s bread and buffer.

        The Pro market is what kept Apple alive during the dark years. Apple relies on the Pro market for its vast influence in the media production industry, via Final Cut, Logic, Ableton, Adobe, etc. They are part of the Mac identity. If Apple is to switch, the Pro market needs to be onboard all the way.

        And there’s the notion that switching Macs to ARM is some kind of “unification” that will lower cost. ARM CPUs for Mac will have to be completely different parts than iOS ARM chips, fab separately form the iOS ones. Plus they have to divide fab to different SKUs because of different Mac in the lineup requires different SKUs. There’s no cost advantage in doing so, it actually increases cost drastically for the CPU part.

      • I agree with Tihomir. This article is suggesting a pipe dream: Apple will not switch to Arm on it’s laptop line. If Apple did, it would only be for the MacBook Air if anything. Everyone talks about “performance that most apps don’t need.” That’s very subjective. End users don’t think that way. They want apps to open and perform as quickly as possible. Also, not all app are created equal. Many applications are developed by developers who don’t know how to optimize; the OS will only do so much.

        I have the 2011 MacBook Air now, and this thing is slow when I open more than 5 tabs in Chrome. And this is using an Intel processor.

        Folks, it ain’t happening unless the performance of ARM is at least remotely in line with Intel. Laptops are different than phones or tablets. Think about it. If I’m a consumer worried about efficiency and power savings, then I’ll just buy a tablet and a keyboard accessory. A Laptop is a different use case all together.