Apple’s Intel-ARM pivot could be a huge Mac moment or blown opportunity

Fifteen years ago, Apple announced that its Mac computers were switching to Intel CPUs, as IBM’s PowerPC desktop and laptop chips had hit unresolvable development roadblocks. At a special media event this month, new Macs will officially replace Intel chips with Apple-developed ARM processors, a transition that could either be a significant milestone for Apple’s computers, or a forgettable moment in personal computing history.

Calling the traditional PC business “stagnant” would be unfair, but a lot has changed since Macs moved from PowerPC to Intel processors. Back in 2005, Macs were Apple’s biggest products, but overall Mac sales weren’t huge — fewer than 5 million units — and Apple’s decade-long chip partner IBM was struggling to create power-efficient parts for next-generation desktop and laptop computers. So then-Apple CEO Steve Jobs warmly embraced Intel, effectively betting the company’s future on its one-time rival. Despite Apple’s modest market share, the PowerPC to Intel transition was one of the decade’s biggest technology stories, cementing Intel as the undisputed leader in personal computer chips, and Apple as a maker of machines enterprises could actually consider buying.

The stakes arguably aren’t as high for Apple today. Switching to Intel chips helped Macs climb from a nadir of 2% global PC market share to a more respectable 10%, but Macs have become far less important to the $2 trillion company’s business over the past decade. In the just-concluded 2020 fiscal year, Macs represented just over 10% of Apple’s sales — roughly one-seventh its total revenues from iPhones, iPads, Apple Watches, and Apple TVs. Despite the strength of its other devices, Apple hasn’t eclipsed Lenovo, HP, or Dell in worldwide PC sales, nor have Macs annually enjoyed even 15% market share in the United States. In theory, Apple could abandon Macs as a legacy business and suffer little financial consequence.

But Apple has kept treating Macs as important to the company’s future, and ultra-niche Mac Pro models as the Lamborghinis of its overall product lineup. To keep winning mainstream customers, Apple has continually streamlined its laptops, multiplied the performance of its desktops, and improved their security with self-developed T-series coprocessors. It has effectively marketed Macs as supersets of Windows PCs, using Boot Camp or emulation to run everything PCs can run (save high-performance games), while offering customers exclusive access to macOS and Mac-only apps.

On a positive note, the Intel-to-ARM transition should be fine for “legacy” Intel-based Mac apps. Apple says they will run properly under Rosetta 2 emulation and — in many though not all cases — have a straightforward Xcode migration path from Intel to ARM instructions. Developers have already had nearly five months to start transitioning their Intel Mac apps to ARM using Mac mini-styled Developer Transition Kits, which Apple has hinted will be blown away by the first Apple Silicon Macs. Given that the PowerPC-to-Intel switch went quite well 15 years ago, most people are predicting that the Apple Silicon transition will be at least equally smooth, at least for macOS users.

Less positively, it’s unclear at this point how Apple will handle the loss of full Windows compatibility that will likely come with a switch to Apple Silicon. The company is throwing out Boot Camp, the tool that let Macs natively run Windows, seemingly leaving virtualization as the only option on post-Intel machines. This means that Windows apps developed for Intel processors could experience performance compromises on Macs, assuming they run at all. Unfortunately, little has been said on this topic by Apple, Parallels, or VMWare, the three companies most likely to offer Windows virtualization, though Microsoft could theoretically cater directly to Mac customers with a performant Windows 10 ARM solution.

A larger question is what advantages new ARM-based Macs will have over “legacy” Intel models. Apple chip lead Johny Srouji offered two explanations at WWDC20 this year: First, Srouji echoed Steve Jobs on the PowerPC-Intel move 15 years earlier, promising that new Macs would gain “a much higher level of performance, while at the same time, consuming less power.” Second, Srouji said that Apple Silicon would include “many custom technologies,” including the following five features:

  • A Secure Enclave for “best in class security,”
  • A high-performance GPU for “a whole new level of graphics performance” on every Mac, suited to both pro apps and games,
  • Neural Engines for machine learning,
  • A video display engine, presumably for screen improvements, and
  • An image processing engine, presumably for camera improvements.

None of those technologies is new — they’ve all been found in iPads and iPhones for years, with some in certain Macs — but if Apple hits the accelerator, they could all represent welcome steps forward for the Mac platform. Even a current iPad Pro-class GPU could up the ante for graphics performance on lower-end Macs that have been saddled with prior-generation Intel integrated graphics chips, while next-generation GPUs could race past Intel’s performance. By contrast, dedicated AI hardware would be a first for the Mac, mirroring the AI cores found in Qualcomm’s Snapdragon 8cx, 8c, and 7c laptop/tablet chips. (Apple didn’t mention 5G modem technology, most likely because it’s not yet ready to be directly woven into Apple’s own chips, but there’s no doubt that integrated 5G will be important to future Macs.)

More consumer-facing changes to the Mac could come immediately from the last two named technologies. Apple could leverage a superior image processing engine to finally bring better FaceTime cameras to the Mac, which has suffered for far too long from mediocre photo and video quality even as iPhones and iPads have continuously improved. This might also mean the addition of Face ID secure unlocking for Macs, a feature that has recently run into challenges due to pandemic-era mask requirements, but would be fine on desktop and laptop machines.

A potential game changer for Macs would be their belated adoption of touchscreen technology, something that Apple dogmatically shunned in Macs for so long that virtually every Windows laptop (and some desktops) beat Macs to the punch. Apple Silicon Macs will be able to run iPhone and iPad apps natively, which is to say that there’s every reason macOS Big Sur should include support for direct touchscreen input, and the first ARM Macs should ship with touchscreen hardware to fully embrace the potential of that compatibility. But it’s entirely possible that Apple once again pooh-poohs touchscreen Macs, leaving them a step behind modern Windows machines in that regard.

There’s no question that the move from Intel to ARM architecture will bring the performance per watt gains Srouji has promised, since Apple’s key chip fabricator TSMC is several nanometer scale generations ahead of Intel in successfully miniaturizing chips, and Apple has been packing its chips with incredible densities of tiny transistors. On the other hand, Apple has been tooting the power-efficient chip horn for a long time, and it’s clear that neither chip efficiency nor resultantly thinner, more beautiful machines have been enough to elevate Macs to the number one, two, or three PC spots globally, or even in Apple’s home country. “Thinnovation” was a nice angle 12 years ago for the MacBook Air, but it’s not clear whether simply making last year’s computers smaller will move the needle on sales, or keep them pretty much where they’ve been stuck for a long time.

Apple’s new chips offer Macs the opportunity to either start a new chapter in their storied history, or instead shuffle forward with the same basic pros and cons they’ve offered for years. We’ll know after the One More Thing event whether Apple Silicon will immediately enable major new functionality, or whether Apple will start slowly by retrofitting older Mac designs with less obvious internal upgrades. In the former scenario, the new Macs will sell themselves, while the latter may require a heaping dose of marketing hype to differentiate Apple Silicon Macs from their Intel predecessors.

Source: Read Full Article