When Apple first announced that it would be transitioning its computers — specifically, the MacBook Air and entry-level 13-inch MacBook Pro, its most popular PCs — to a new and wildly different type of processor, there were plenty of reasons to be skeptical. Apple was making huge claims for battery life and performance, things that the first wave of Arm-based laptops from Qualcomm and Microsoft failed to deliver.
But deliver Apple did, with computers powered by a new M1 processor that aren’t just close to their previous Intel counterparts, but crush them in nearly every respect — and not just the base model Intel chips that the M1 purports to replace, either. In both early benchmarks and head-to-head comparisons for compiling code, Apple’s M1 chip appears to hold its own against even Intel’s most powerful Core i9 chip for laptops.
Keep in mind this comparison is deeply unfair: my 16-inch MacBook Pro was literally maxed out just a year ago – 8 cores, 64GB RAM, and much more, costing $6000.
In comparison the M1 costs just $2000 and manages to hammer the Intel machine with a quarter of the RAM.
— Paul Hudson (@twostraws) November 17, 2020
The conversation has flipped instantly: it’s no longer “why would you take a gamble on Apple’s new, unproven processor” but “how will competitions like Intel, AMD, and Qualcomm respond?”
For years, Intel and AMD have been playing a chess match, sniping back and forth with improvements in CPU performance, battery life, and onboard graphics. Apple appears to be playing an entirely different game on an entirely different level. The same interplay between hardware and software that has led to such huge successes on the iPhone and iPad has now come to the Mac.
It’s not just that Apple’s hardware is faster (although straight benchmarks would indicate that it is); it’s that Apple’s software is designed to make the most of that hardware, in a way that even the best optimization of macOS on an x86 system wasn’t. As John Gruber notes (citing Apple engineer David Smith) the new chips handle fundamental low-level macOS app tasks up to five times faster on the M1 than they do on Intel because Apple was able to design a chip from the ground up to specifically be good at those tasks. It’s why the new M1 Macs (and the existing iPhone and iPad lineups) are able to do more with comparatively less RAM than their Intel (and Android) counterparts.
Apple has also done incredible work with Rosetta 2, its translation layer for running legacy x86 applications on the M1. It’s another key part of how Apple’s software strategy pays off big dividends for the new hardware by making it seamless to run older software on the new Mac without any real hits to performance. Apple almost certainly has factored Rosetta 2 optimization into the M1’s design, benefiting from the same parallel development as the rest of the hardware. The result is that M1 laptops don’t make users choose between great performance on Arm-optimized apps at the expense of legacy x86 performance; instead, they run old apps well and new optimized apps even better.
The most exciting — or frightening, if you’re a traditional PC chip company — part of Apple’s new chips is that the M1 is just the starting point. It’s Apple’s first-generation processor, designed to replace the chips in Apple’s weakest, cheapest laptops and desktops. Imagine what Apple’s laptops might do if the company can replicate that success on its high-end laptops and desktops or after a few more years of maturation for the M-series lineup.
Right now, the saving grace for traditional x86 laptops is that it’s only Apple, with its near-complete control over its hardware and software stack, that’s managed to accomplish this level of speed, software performance, and battery life on Arm.
It’s an open question whether companies like Qualcomm and Microsoft will be able to emulate Apple’s success with the next wave of Arm-based Windows machines. Certainly, it would take a much bigger restructuring of Windows, one that would impact a far greater number of customers than Apple’s changes. And while Microsoft does design its own Surface laptops — and even worked with Qualcomm on building Arm-based SQ1 and SQ2 chips for its Surface Pro X lineup — it’s still a far cry from the level of control that Apple maintains over its software / hardware ecosystem that allows so much of the M1’s success.
The new MacBook Air and MacBook Pro won’t be the perfect laptops for everyone, especially if you rely on huge, GPU-intensive tasks or specific developer tools. But when a $1,000 M1 laptop can outdo a maxed-out, $6,000 MacBook Pro with quadruple the RAM and Intel’s best chip, while also running cooler and quieter in a smaller and lighter form factor and with twice the battery life, where do competitors even go from here?