It's not a new rumor, but it's gained a new head of steam in recent days: Apple is designing its own chip to power future Macs, in place of the Intel chips it has used since 2006.
Apple already designs its own A-series chips using the ARM architecture for the iPad and iPhone, and those chips have outpaced other ARM-based chips. They've been 64-bit, for example, for more than a year. Apple clearly knows what it's doing in chips.
But a new chip means existing apps won't run as is, even if OS X is ported — a significant barrier to switching the processor in the Mac. Apple has done this before — twice, in fact. First, it switched in 1991 from the Motorola 68000 series to the IBM PowerPC, then again in 2006 from the PowerPC to the Intel x86. In both cases, it created compatibility modes, largely invisible to users, that let old apps still run on the new chips. Apple gave developers several years to port their apps.
When Apple adopted the x86, it also gained the ability to run Windows on Macs both directly in a Boot Camp partition and indirectly via a virtual machine from OS X. That provided a safety valve for people switching from Windows to OS X, but the safety net is not nearly as useful today as it was in the 2000s, given the rise of Web apps that make the chip less of an issue and the increase in Mac-compatible apps.
Still, for some, that safety net is critical, and it's unclear whether Apple could still support Windows on an ARM chip. (I'll call that chip the B-series, to distinguish from the A-series used by iPhones, iPads, and iPod Touches.)
However, enough has changed that abandoning the x86 is more plausible today than it was even three years ago. Apple has a unified IDE, Xcode, for OS X and iOS, and I have no doubt that Apple would have a future Xcode generate the code for existing apps to run on a future B-series Mac chip.
Perhaps more important, the iPad's success means there are already a lot of ARM-compatible apps that could run on a B-series Mac, in addition to or instead of x86 Mac apps. Much of Apple's own software — Mail, Calendar, Notes, Contacts, Messages, FaceTime, Safari, the iWork suite, GarageBand, iMovie, and the forthcoming version of Photos — have strong ARM versions as well as Intel iterations, so they could largely slide in place of the Intel apps on a B-series Mac.
But other Apple apps — notably iTunes, Preview, Image Capture, Time Machine, TextEdit, iBooks Author, Final Cut Pro, Logic Pro, OS X Server, Configurator, iTunes Producer, and of course Xcode — would need to be rewritten for the B-series Macs.
Microsoft's powerful Office suite for iOS could satisfy most Office users' needs on a B-series Mac, too. Microsoft's OWA app is a pale imitation of Outlook, but at least there's a core Microsoft could build on to make an equivalent B-series app.
Google has likewise ported many key apps to ARM, including Chrome, the Google Apps suite, Gmail, Google Drive, Google Hangouts, and Google Now. Few of those are available for Intel Macs or PCs anyhow outside of a Web browser.
But as you get into more specialty apps, the app gap widens. Adobe's Creative Suite has no ARM versions, save a very limited versions of Illustrator, Lightroom, and Photoshop. AutoCAD has a limited ARM-based version of AutoCAD for the iPad. The list largely stops there.
People who use such specialty apps — for CAD, visualization, creative design, and video editing — are the ones who buy the $2,500 iMac 27-inch computers and $3,500 Mac Pros that Apple's PC reputation is based on. It's hard to see such apps running well on a B-series chip under emulation, and getting their developers to rework the apps is asking a lot for what remains a small market.
Jean-Louis Gassée, former head of Apple France and now an industry analyst, doesn't see any reason for Apple to abandon Intel on the Mac and create this app gap. There's no cry for a unified OS across Macs and iOS devices, he notes, and the higher cost of the Intel chips likely wouldn't justify the cost of porting all those apps to the new B-series chip. "What job is to be done?" he asks.
I don't have an answer to that question, and if Apple doesn't either, then this switch won't happen. Still, if Apple has a reason, it's perfectly capable of making such a switch happen. The rise of iPad apps would make the app-porting effort easier than many people realize.