It happened again: a report that Apple will soon ditch Intel chips for its Mac lineup. But while that rumor pops up as dependably as a perennial, this time it has enough weight to merit some serious consideration. And the first thing to consider is just how hard it would be for Apple to actually pull this off.
It seems likely the company will give it an earnest effort. The report Monday comes from Bloomberg Businessweek’s Mark Gurman, who has the best track record of any Apple outsider in divining Cupertino’s plans. And in some ways, the company has already spent the last few years laying the groundwork, not only investing heavily in its own homegrown processors, but also hewing together elements of its iOS mobile operating system and MacOS desktop counterpart that would make this kind of switch feasible.
Still, quitting Intel would come with a host of complications. How Apple navigates them will shape its next decade and beyond.
Intel has provided Apple with processors for its Mac lineup since 2006, a long and mutually profitable relationship. While MacBooks and iMacs lack the marquee appeal of the iPhone, Apple still sold nearly $7 billion worth of Intel-powered laptops and desktops just last quarter; Apple reportedly provided about four percent of Intel’s revenue last year.
Which seems pretty reasonable, as far as symbiosis goes. And yet Apple’s desire to go it alone should come as no surprise. The company already, after all, makes its own A-series processors for not just the iPhone but also the S-series for the Apple Watch, the W-series for wireless headphones, as well as coprocessors that have already found their way into the Mac line. It even, as of a year ago, is designing its own GPU.
At this point, then, Intel’s presence in the Mac lineup is more the exception than the rule. And the reason Apple would want to create those chips in-house is the same reason major smartphone makers have been ditching Qualcomm: If you can do it yourself, don’t rely on someone else.
“It all seems to be about going deeper and deeper on what they’ve demonstrated, which is that the more of the technology stack that you control, the more you can refine it,” says Frank Gillette, analyst at Forrester Research. Designing the A-series lets Apple customize the chip specifically for the iPhone’s needs. The W-series gives the Air Pods more robust Bluetooth powers. Even the iMac Pro has a secondary, Apple-made T2 chip that provides a security boost. Deploying that same strategy across its traditional computers would similarly help Apple stand out. It would also mean Apple could release new products at its own cadence, rather than hitching its wagon to Intel’s sometimes inconsistent star. Intel declined to comment.
All of which would seemingly make Apple’s transition away from Intel not just likely, but inevitable. As anyone who’s tackled Queen at karaoke can tell you, though, there’s a huge gap between wanting to accomplish something and succeeding.
As Gurman reports, Apple hopes to replace the x86 Intel architecture that its Macs have used for over a decade with ARM-based chips, like those that power the iPhone. That transition would pose at least two hurdles, both fairly high.
The first is the processor itself. ARM-based designs are valued for their efficiency, but in terms of pure horsepower, they can’t currently come anywhere close to matching Intel’s upper-tier guts. And while Apple’s not expected to make the switch until at least 2020, observers doubt that’s enough time to catch up.
“Computationally I can see a Core i3 or low-end Core i5,” says Patrick Moorhead, founder of Moor Insights & Strategy, comparing ARM’s abilities to entry-level Intel chips. “I can't imagine that by 2020 they’d have a processor anywhere near the capabilities of a Xeon or a Core i7.”
There are some potential solutions there. Apple could simply transition its entry-level MacBook to an ARM processor and leave Intel in its pro-focused lines until ARM catches up to their needs as well. And Apple could, over the coming years, shift some of the CPU’s traditional responsibilities to the GPU, which it already controls.
“We’re already integrating GPUs to be able to handle a lot of those tasks,” says Eric Hanselman, chief analyst at 451 Research. “When GPU acceleration becomes one of the core elements of tackling more complex computational problems, ARM architectures actually start making a lot more sense, because you now have the ability to integrate and build customized environments much more easily.”
Segmenting its lineup seems like the likeliest path; Gurman describes a "multi-step transition." But that would also create potential headaches of its own, both for developers and consumers. Apple is reportedly working on a platform that allows developers to write the same app for both MacOS and iOS, but that sort of hybrid invites complications, especially if some devices switch to ARM and others remain on Intel.
“All of the Intel products are on MacOS, not iOS,” says Moorhead. “MacOS and iOS are very different, in terms of multitasking, the amount of threads that should be used, peripheral support. You can plug almost anything you want into a Mac. You’d have to enable that into iOS.”
Developers, meanwhile, might have their work cut out for them reworking their apps for an ARM-based version of OS X, just like they did when Apple transitioned to Intel over a decade ago.
“There’s no magical way that Apple can make that complexity disappear. Just like we saw when Apple went from PowerPC to Intel, there wasn’t some magic way to make PowerPC apps on the Mac just work on Intel,” says Moorhead. “Most of them had to be recompiled. A lot of them had to be rewritten.”
Apple could also find users flummoxed at its attempt at the MacOS-iOS mashup that would apparently accompany an ARM transition. It wasn’t so long ago, after all, that Microsoft flamed out spectacularly when it attempted to bring a mobile UI to the desktop in Windows 8, an overhaul that left users feeling mostly confused and annoyed. And while Cupertino has already made some adjustments to give its desktop and mobile operating systems some common ground—its Apple File System, introduced last spring, works across both—it will have to combat years of ingrained expectations about how Apple devices behave.
“The operating system challenge looks like a significant one to me, especially given that they’ve kept the graphical user interface approach and the touch approach completely separate,” says Gillette. “I think quite honestly they face a huge technical channel on the processor, and a huge emotional challenge for the end user.”
And in the meantime, Mac developers have little incentive to put significant work into their applications between now and 2020. And potential Mac buyers have every reason to sit on the sidelines until then. Which means that significant change will likely be preceded by a crippling stasis.
All of these potential pitfalls are surmountable, especially given a long enough time horizon. And the potential benefits to Apple seem worth it.
“I’m sure the attractiveness of the ARM licensing model is something that lets them think that this ought to be something that would unbind them from the normal Intel cadence, would give them more control, and would let them have an environment in which they manage much more of not only their capabilities, but also the intellectual property that sits throughout this whole stack,” says Hanselman.
But breaking up with Intel won’t be as easy as flipping a switch. If it were, Apple would have done it years ago. It’s going to require risk from Cupertino, hard work from developers, and a willingness from MacBook users to rethink how their devices work. And that’s before you get to the professional-grade users, and when and how an ARM-based processor might accommodate their needs.
It’s clear why Apple has charted a path toward homegrown Mac processors. But the move away from Intel itself isn’t the surprising or interesting part of this. It’s how it will shape that future without breaking it.
- A year ago, Apple announced that it would make its own GPU to control its own destiny
- Apple put a "neural engine" in the A11 processor to help handle complex artificial intelligence tasks
- Major smartphone makers have already made the jump to in-house processors. Why not PC makers too?
