On Monday Bloomberg News dropped a bombshell report. By 2020, it claimed Apple will stop using Intel CPUs in its computers. Just picture it: The third-largest PC maker in the world might one day leave behind the biggest computer chip maker.
Quickly, Intel’s stock dropped following the news, and while much of Twitter was aflutter with gleeful takes, others were more wary—and even worried. It would be relatively unprecedented for Apple to make its own CPU chips, and the move could mean a variety of things. So let’s dive into what is actually happening and what it all could mean.
Apple and Intel breakup rumors are fairly common, as Patrick Moorhead, president and principal analyst for Moor Insights & Strategy—with an expertise in CPUs, pointed out on Twitter yesterday.
It seems like every time Apple and Intel come to the negotiation table the rumors flare up—with the press being used as a tool in the process, as Moorhead noted. In the last two years alone similar rumors have popped up on at least four separate occasions.
The difference here, however, is the source of the latest rumor. Bloomberg writer Mark Gurman has a history of securing leaks from Apple, a company notorious for its secrecy.
Theoretically this could just be Apple using a reporter to play hardball in negotiations, but let’s presume Gurman isn’t Apple’s pawn and this information is legitimate. According to Gurman’s piece, co-written with Bloomberg reporter Ian King, Apple has a new initiative internally known as Kalamata, and the goal of Kalamata is to put Apple-designed and made chips in Macs by 2020.
Everything beyond that rumor is further supposition on the part of reporters and commentators writing about the rumor.
But since Bloomberg broke this story, the suppositions and theories for what could happen have been flying around, so let’s try to get a better understanding of the speculation.
This is absolutely the most popular theory available, and it has been floated by both Bloomberg and the Verge. In this scenario, Apple doesn’t just say goodbye to Intel; it says goodbye to the x86 architecture that Intel chips operate on, and instead embraces ARM, which is what the CPUs in the iPad and iPhone are based on.
The theory is that Apple is very good at making super fast and efficient mobile phone CPUs and that Intel has recently stagnated—failing to come up with a major, and necessary, CPU redesign in over three years. So Apple, which despite its size and success in the mobile space still has no measurable background in computer processors, could theoretically do what the largest producer of computer processors in the world cannot.
Now, if anyone can succeed here it is certainly Apple. What the company lacks in computer processor know-how, it makes up for in resources. Heck, the company could probably buy Intel or its rival AMD and still have money left over to buy some other company everyone thinks Apple should buy.
But it would, by no means, be easy. Robert Graham of Errata Security wrote an entire Twitter thread yesterday about the problems of using ARM processors for the kind of high-end computing that people buy iMac Pros and MacBook Pros for.
In the thread, Graham notes that ARM processors are known for their power efficiency because they’re not actually intended to run as fast as the x86 processors found in nearly every major laptop, desktop, and server. It might work well for a Chromebook or an iPad, or the very specific needs of Cloudflare’s server farms, but it wouldn’t necessarily translate well to the demands of someone rendering a few hours of 4K footage or even a few minutes of stuff produced in Adobe After Effects.
There’s also another big problem: macOS was designed to work on the x86 architecture. Moving to ARM would require a major redesign of how the software works and that means eventually support for x86 computers would end, which is exactly what happened the last time Apple made a major architecture change in 2006. Back then, it moved from Motorola’s PowerPC to Intel. The last operating system to support PowerPC was Mac OS X Leopard, which launched in October 2007—just a year after the move began.
Besides the end of legacy support for macOS, the shift from x86 to ARM would also affect software developers.
The move would likely be relatively seamless for the sort of developers who operate primarily in Apple’s own software languages. But it could hit the big software guys harder. Companies like Adobe often use code to take advantage of the CPUs and even the GPUs of devices. Their software is currently optimized for the x86 architecture and could need some significant editing to work as well on ARM.
And if the you think those changes to the software could come quickly think again. Companies like Adobe, Microsoft, and Autodesk are all already struggling to provide the exact same experience on Mac and Windows—throwing in an architecture change could make that more difficult.
Just look at Google’s Chrome OS. It is enormously popular in the education sector, but its users still can’t access Adobe’s full-fledged desktop software, nor can iPad users, though Apple continues to push the tablet as a Chromebook competitor.
Also look at the troubled history between Apple and Adobe. Besides fighting over Flash on iPhones, the two companies have bickered about Adobe support in Mac OS X, and at one point, in 2003, Adobe even completely pulled support for Premiere because it would have taken more work than the company thought was financially viable.
The growing pains for an all-out move to ARM would likely be severe, and the potential to alienate the market that kept Apple afloat for a couple of decades—the creative market—would be extreme. If Apple’s ARM processors couldn’t handle the heavy workloads required by creatives they’d stop investing in Apple products and that could lead software companies to stop investing, too.
But if there’s a company that can browbeat us all into loving a new macOS that’s more in line with iOS and features hamstrung apps not as capable or as quick as their predecessors (on machines slower than anything with Windows on it), it’s Apple.
Only maybe twisting everyone’s arm and forcing them into a new ecosystem isn’t the actual plan—though Bloomberg’s Gurman has repeatedly reported on Apple’s plans to further integrate Apple hardware and software and move towards some kind of more unified OS. Maybe instead of one big shift Apple has some kind of bridge device planned? Marrying two worlds.
To continue catering to the high-power demands of, say, video editors and illustrators, Apple could decide not to move all its computers off of x86. Instead, it could only move its cheaper computers to ARM. That would mean either a whole new MacBook or MacBook Air—the two least expensive Apple computers currently available—or something brand new, as suggested by former Wall Street Journal columnist and the Verge and Recode editor Walt Mossberg.
In many ways, this feels more viable. Instead of trying to twist ARM into something it isn’t, or radically redesigning macOS and demanding major software partners move to designs for a new architecture, Apple would get its cheap laptop based on its own chip that would also happen to get people more inured to the idea of macOS and iOS merging one day far in the future. A new machine to herald Apple’s new ecosystem.
In the short term, this kind of device seems highly likely, but we have no idea what the OS would look like. iOS would make sense, as it’s already designed to run on ARM processors. Trying to force macOS onto it would make less sense. Earlier this year, Microsoft partnered with Qualcomm, which makes mobile CPUs based on ARM, and Asus, which makes laptops, to release a Windows 10 laptop with a Qualcomm Snapdragon processor. The battery life was incredible—the speed was significantly less so.
Putting macOS, in its current state, on such a device, even with Apple’s great CPUs inside, would likely not end well. But maybe Apple wouldn’t actually embrace ARM as the architecture for all its devices.
In this scenario Apple stops using Intel CPUs, but continues to use the x86 architecture. This would mean no big fundamental changes to macOS or your favorite software, and no threat of losing powerful computers like the iMac Pro. Ideal right?
Only, as far as we know, Apple isn’t equipped to start designing x86 CPUs. It’s made no publicly known major acquisitions or licenses for intellectual property that would suggest it has the team to build those chips, and there does not appear to be any in-house designers there who are particularly known for their work in the x86 space.
They’d really need a pal with x86 know-how to make it work.
This is the true moonshot scenario; Apple stops buying Intel CPUs and instead starts working with AMD on semi-custom chips based on the x86 architecture.
For people who don’t want Apple to move away from x86, this makes a lot of sense. AMD has a very aggressive roadmap for the release of CPUs and GPUs over the next few years, its current CPUs and GPUs have been met with praise, and it already licenses its tech to others, including Sony (for the Playstation 4), Microsoft (for the Xbox One), and Intel (for the new G-series of processors).
Apple could do a deal with AMD and start producing wicked fast chips in decently short order. But this, and any scenario that includes keeping the x86 architecture, seems to fly in the face of the rumors that Apple wants to unify its software and hardware. Joining forces with AMD might help Apple in the short term, and could definitely give it a better understanding of how to design CPUs for laptops, desktops, and servers, but it wouldn’t necessarily move Apple towards any kind of grand unification.
What Apple will actually do remains to be seen. But nearly any way you look at it, the CPU space is about to get a lot more interesting.