How 30 years of chip transitions paved the way for the spectacular Apple Silicon era

With the announcement that macOS Tahoe will be the last Mac OS version to support Intel Macs, Apple’s preparing to close the books on the third chip transition in Mac history.

It doesn’t get a lot of attention, but Apple is absolutely the best company in the world at picking up stakes and moving its platforms somewhere else. Over its 41 years of existence, the Mac has run on four entirely different processor architectures (not to mention two different operating system foundations), all the while remaining more or less the same familiar Mac we know and love.

This is not an easy feat to accomplish once, let alone three times. Apple’s gotten very good at this. Twenty years ago, it was the switch to Intel. Five years ago, the switch to Apple silicon started. And of course, way back in the mists of time when I was a brand-new hire at one of Macworld’s predecessor publications, Apple made the leap for the very first time.

Hard-earned lessons

When I wrote about the Mac’s history of chip transitions as rumors of the Apple silicon transition swirled, I made great pains to point out that Apple had been there, done that, and learned many lessons. My only concern was if there was anyone left at Apple who had lived through the old transitions, or if the company was going to have to figure it out from scratch all over again.

After that story was published, I heard from someone inside Apple who assured me that, yes, there were still some people kicking around who had been there when that very first chip transition happened back in the 1990s. That kind of institutional memory was vitally important to Apple’s transitions in 2005 and 2020, as it turned out.

The original Mac came with a Motorola 68000 processor. The 68K was used on all sorts of video games, some Atari computers, as well as the Mac. But in the early 1990s, Apple was frustrated by the slow pace of improvements by its chipmaker and realized that the fate of its platform was dependent on the success or failure of someone else.

This story will keep repeating.

Apple was doing its own chip research in those days when, in the most unlikely of events, its arch-rival IBM approached the company about collaborating on a next-generation chip design. With almost-jilted Motorola added in as a partner, the AIM Alliance set about building a new set of chips that would become PowerPC.

Foundry

PowerPC was a next-generation chip with features that differentiated it from the dominant Intel processors of the day. The first Macs with PowerPC chips inside, dubbed Power Macs (of course!), arrived in March 1994. To get to that point, Apple didn’t just have to bring its software to a new chip design; it had to enable compatibility with old Mac software.

Power Macs ran a (nameless) 68000-series emulator that enabled them to run non-native software at a slight speed penalty. My memory is that the sainted Microsoft Word version 5.1, which was not PowerPC native but was awesome, was still quite usable (though noticeably slower at some tasks) on the new chips.

As a first-time transition with a base of users who had committed to the Mac during its first decade, it was a scary time. For a year, we ran a column called “Ask Dr. Power Mac,” where users wrote in to understand the technical challenges they might face as they upgraded.

Apple’s biggest mistake in this era was that it didn’t control its developer tools. Metrowerks, a software company ultimately bought by Motorola, built the definitive PowerPC development environment, CodeWarrior. (Apple would learn a key lesson from this; today, pretty much all development happens in Apple’s own Xcode.)

Within a year, the transition picked up speed, new PowerPC native software shipped, and Apple had a template for any future processor transitions that the Mac might require. (But hopefully not, right?)

Everything falls apart

It’s the summer of 2003, and so far as anyone knows, the PowerPC era is proceeding apace. The new G5 (fifth-generation) processor has been announced, and Steve Jobs has promised that it’ll eventually reach an all-time record 3GHz speed. The community is excited about the promise of that power coming to Mac laptops as well. During the East Coast Macworld Expo that year, Apple PR proudly takes me on a tour of IBM’s chip plant in Fishkill, New York, where the cutting-edge G5 will be produced.

That was a major turning point, but not in the way Apple intended. IBM was never able to produce that 3GHz chip for Apple. The G5 wasn’t suitable for laptops. And deep inside Apple, a skunkworks project was making sure that the brand-new Mac OS X could run on Intel processors. Twenty years ago, Jobs announced the switch on stage at WWDC: The AIM fellowship was severed, and Apple would move from PowerPC to Intel.

This time, Apple gave the technology that translated PowerPC code and ran it on Intel processors a name: Rosetta. The PowerPC software it emulated ran slower, to be sure, but Intel-native “universal” apps appeared quickly, and faster Intel processors kept appearing at a rapid pace. The Mac had never been faster, and perhaps more importantly, could no longer be negatively compared to the speed of Windows PCs.

This period was, in many ways, the most important decade in the Mac’s history. The growing success of the iPod (and later, the iPhone) put the Mac in front of people who might never before have considered buying one. A new generation of Windows emulators, capable of running at full speed on Intel hardware, provided a fallback for PC users who might need to run a handful of Windows programs. The Mac began to grow rapidly.

Doing it for themselves

It was good while it lasted, but 15 years after the Intel era began, Apple turned the page. Once again, the company was frustrated by the pace of chip development and its lack of control over one of its platforms. But there was a key difference this time: Apple had, for a decade, been designing its chips for the iPhone and iPad. Developers were building apps in Xcode that compiled and ran on Apple’s processors.

This was, in many ways, the easiest chip transition Apple has made. The tools were there. The developers were familiar with Apple’s chips. Apple had years of experience to give it the confidence that it could apply what it had learned building iPhone and iPad chips to create powerful Mac variants.

The results were immediate: The M1 Macs, when they arrived in the fall of 2020, were the best-reviewed Macs in recent memory. They were so much faster than their Intel predecessors that, in some cases, Rosetta 2, the latest version of Apple’s code-translation layer, ran Intel apps faster than they did on the original Intel hardware.

The rise of the web and mobile platforms meant that Windows compatibility wasn’t as important as it was in 2005. And in a pretty amusing wrinkle, Microsoft had already begun its own weird sort of chip transition, building a version of Windows that ran on processors very similar to those being used by Apple. (It’s even got its own code-translation layer. Clearly, Microsoft learned from the best.)

Which brings us to the final question: If Apple has changed its Mac chip architecture after 10, 11, and 15 years, does that mean the Apple silicon era will also end?

Anything’s possible, especially in the tech industry–but the big difference is that now Apple designs its own chips, to its own specifications, in tandem with the products it builds. That’s a huge advantage that it’s never had before.

Of course, Apple thought the same when it formed the AIM alliance. And when it tied itself to Intel, which was the world’s dominant chipmaker at the start of that partnership but had been supplanted by TSMC by the time it ended. Life comes at you fast. But, at least for now, the Mac endures as the world and chips change around it.