Why Apple is not about to release ARM-powered Macs

If someone had said to you last Christmas that the UK would decide to leave the EU, and Donald Trump would be a serious challenge to Hillary Clinton in the US Presidential Election, you’d probably have laughed at them. So why doesn’t one more wild story come true, that Apple is about to dump Intel processors and chipsets from Macs, and switch to ARM?

There are many sites – including iClarified, TechnoBuffalo, MacTrast, and MacDailyNews – now claiming that, at its event on 27 October, Apple will announce new Macs which will use its own integrated systems on a chip (SoC) based on the ARM Hurricane processor, as used in the iPhone 7 and 7 Plus, or perhaps just ARM Hurricane cores.

There are a few small problems with this story before it should attract any credibility.

The ‘evidence’ on which this story is based is not actually new. macOS and iOS share a lot of code, and one way which Apple uses to make this easier is to have common sourcecode files, with conditional content. Look back as far as OS X 10.7 and similar ‘evidence’ appears there. This new reference to ARM Hurricane support simply means that macOS Sierra has common source with iOS 10, which we all knew anyway.

The last time that Apple changed processor architecture in Macs, from PowerPC to Intel, Apple announced that change in June 2005, seeded developers with prototype Intel Macs soon afterwards, and released its first consumer models in January 2006. Major hardware changes, such as switching processor architecture, are far more disruptive to developers and Apple’s own engineering teams than a mere major version upgrade to macOS/OS X, which we’re still going through.

Before anything else can happen, Apple has to release a complete new version of its development system, Xcode, and test all its tools, such as compilers, and the many software libraries to support the new architecture. Major vendors like Microsoft and Adobe who have used their own in-house development systems have to port those too, test, and bring them to release quality before they can even start the job of moving major apps across to new processors.

Unlike iOS devices, a lot of peripherals and other products for macOS depend on low-level software including drivers and other kernel extensions, which are often hardware-specific. Suddenly releasing new Macs which have a different processor architecture would ensure that hardly any peripherals – even basics such as graphics tablets and hardware RAID drives – would work with those new models for many months. Given the worrying number of printers, scanners, and other devices which are still unsupported in Sierra, this would make those new Macs almost useless.

One of the great success stories in Apple’s transition from PowerPC to Intel processors was its PowerPC software emulator Rosetta, which enabled a lot of existing products to run, albeit rather slowly, on its new Intel Macs. Rosetta had serious limitations, and could not, for example, run kernel extensions, but most significantly apps running on Rosetta were slow. If Apple were to release new Macs which actually ran all existing software more slowly than old ones, they would flop.

Finally comes the issue of timing. Apple has announced significant hardware changes first and foremost to developers, at WWDC, because it knows that they have to be won over to support the new hardware with their products. Going into the lucrative 2016 Christmas season with a range of new Macs which are incompatible with almost every peripheral available, and run existing software more slowly than old Macs, would be commercial suicide.

But it has been a very strange year, when everything predicted by common sense seems to have gone out of the window. Despite that, I am confident that we will see new Macs, but that they will still be based on Intel CPUs and their supporting chipsets.