Apple M1 — The Mac Moves to Apple's Own ARM Chip

Apple announced the M1 chip on 10 November 2020. The same day it opened orders for the first three Macs to use it — the MacBook Air, the 13-inch MacBook Pro and the Mac mini — and they reached stores and customers on Tuesday 17 November 2020. Japan got them on the same day, 17 November, so it was in the first wave.

US prices started at $999 for the MacBook Air, $1,299 for the 13-inch MacBook Pro and $699 for the Mac mini. In Japan they started at ¥104,800, ¥134,800 and ¥72,800 respectively, excluding tax. All three looked almost exactly like the models they replaced. What changed was the chip inside.

How the M1 came about: the timeline

In June Apple had said it would keep releasing new versions of macOS for Intel-based Macs for years to come, and that it "has exciting new Intel-based Macs in development". Sure enough, an Intel 27-inch iMac followed in August. November 2020 was not the end of the Intel Mac but the beginning of the end. Against Apple's forecast of about two years, the completion it declared came with the Mac Pro roughly three years later.

What was inside the M1

The M1 was the first chip Apple designed specifically for the Mac, combining the CPU, GPU, Neural Engine and more on one system on a chip (SoC).

Apple claimed the world's fastest CPU core in low-power silicon and the world's fastest integrated graphics in a personal computer. Both are Apple's claims, based on its own testing. The chip is generally reported to be made by TSMC, but Apple's press release does not name the manufacturer.

In Apple's description, unified memory let every part of the SoC work on the same data without copying it between separate pools of memory. M1 Macs came with 8GB or 16GB.

Reading "up to 3.5x"

Apple said the M1 delivered up to 3.5x faster CPU performance, up to 6x faster GPU performance, up to 15x faster machine learning and up to twice the battery life of previous-generation Macs. Those headline numbers gather the biggest gains across the three machines. Apple's per-model claims, each against the model it replaced, were:

All are Apple's measurements, not independent results.

How Intel apps kept running

Changing architecture means existing apps do not run as they are. Apple bridged the gap two ways: Rosetta 2, which translates apps built for Intel, and Universal 2, which packs Intel and Apple silicon code into one app. M1 Macs could also run iPhone and iPad apps directly. All three shipped with macOS Big Sur.

Apple had done this before. In the move from PowerPC to Intel, announced in June 2005 and begun in 2006, the first Rosetta let PowerPC apps run on Intel Macs; Rosetta 2 takes its name from it (see also ARM vs x86).

ARM elsewhere in 2020

The M1 arrived as ARM chips were spreading beyond phones.

• AWS Graviton2: on 11 May 2020 Amazon made EC2 instances on Graviton2, its own chip using 64-bit Arm Neoverse cores, generally available, claiming up to 40% better price performance than current x86-based instances
• Surface Pro X: in 2019 Microsoft had released the Surface Pro X, running Windows on ARM on its SQ1 chip, based on Qualcomm technology
• NVIDIA's bid for Arm: on 13 September 2020 NVIDIA announced it would buy Arm for $40 billion. It abandoned the deal on 7 February 2022, citing significant regulatory challenges

The ARM architecture, born in 1985, had long belonged to phones and embedded devices. Across 2019 and 2020 it moved into places x86 had held — cloud servers, Windows laptops and then the Mac.

What followed

In 2021 the M1 also went into the iMac and iPad Pro, and that October the larger M1 Pro and M1 Max arrived. Every Mac ran on Apple silicon from the Mac Pro of June 2023.

The companies are on Apple and Intel; other hardware milestones are on the hardware timeline.