The first Macs built on Apple's own M1 chip went on sale in November 2020: the MacBook Air, the 13-inch MacBook Pro and the Mac mini. The M1 is an Arm-based system on a chip (SoC) made by TSMC on its 5 nm process, and it replaced Intel processors in those models. It brought a big jump in performance per watt, which is why the first M1 laptops ran cooler and much longer on battery than the Intel models they replaced.
This is a news post first published on 8 December 2020. Statements such as "fastest" and "most powerful" describe the chip at launch.
Key takeaways
- Apple announced the move from Intel to its own Apple silicon at WWDC in June 2020 and unveiled the first M1 Macs on 10 November 2020.
- The M1 is closely related to the A14 Bionic used in the iPhone 12 and iPad Air, and is built on TSMC's 5 nm process with about 16 billion transistors.
- It has an 8-core CPU (four performance and four efficiency cores), up to an 8-core GPU, a 16-core Neural Engine and unified memory on the package.
- At launch it gave the Mac its fastest single-core performance and beat many Intel laptop chips at a fraction of the power.
- Apple later extended the family (M1 Pro, Max and Ultra) and moved on to M2 and later generations.
The transition from Intel to Apple silicon
At WWDC 2020, Apple announced plans to move the Mac away from Intel chips to Macs built with its own Apple silicon, starting in late 2020. Apple's custom chips are Arm-based and similar to the A-series chips used in iPhones and iPads. Apple unveiled the first Apple silicon Macs in November 2020 and said the full transition would take about two years.
This was the Mac's third processor change, after the moves from Motorola 68000 to PowerPC in 1994 and from PowerPC to Intel in 2006. As before, Apple provided a translation layer, Rosetta 2, so that existing Intel software could run on the new machines while developers rebuilt their apps for Arm.
What is inside the M1
The M1 is Apple's first system on a chip for the Mac. It integrates many components that used to be separate chips on an Intel Mac's logic board: the CPU, GPU, unified memory, Neural Engine, Secure Enclave, SSD controller, image signal processor, media encode and decode engines, and a Thunderbolt controller with USB 4 support.
| Block | M1 at launch | What it does |
|---|---|---|
| CPU | 8 cores: 4 performance (Firestorm), 4 efficiency (Icestorm) | General computing; heavy single-threaded work on performance cores, light tasks on efficiency cores |
| GPU | 7 or 8 cores, depending on model | Graphics and GPU compute |
| Neural Engine | 16 cores, rated by Apple at 11 trillion operations per second | Machine learning tasks such as image and speech processing |
| Unified memory | 8 GB or 16 GB on the package | One memory pool shared by CPU, GPU and Neural Engine |
| Process | TSMC 5 nm, about 16 billion transistors | Manufacturing technology |
| I/O | Thunderbolt and USB 4 support | External displays, storage and accessories |
Performance and efficiency
At launch, the M1 was the most powerful chip Apple had made for the Mac. It is similar to the A14 chip in the iPhone 12 and the iPad Air of that time, and both are built on a 5-nanometre process by Taiwan Semiconductor Manufacturing Company (TSMC).
The 8-core CPU pairs four high-performance cores with four high-efficiency cores. The performance cores handle demanding, single-threaded work, while the efficiency cores handle lighter tasks such as web browsing at much lower power. This split, together with the 5 nm process, is a large part of why M1 Macs have such long battery life. Apple quoted up to 18 hours of video playback for the M1 MacBook Air and up to 20 hours for the 13-inch MacBook Pro.
The 8-core GPU and 16-core Neural Engine complete the chip. Compared with Intel Macs, the M1 gave the fastest single-core performance of any Mac at the time, and independent reviews found it ahead of the Intel chips in the 16-inch MacBook Pro in many tasks, with performance close to many desktop machines.
Why an SoC with unified memory matters
On a typical PC of that era, the CPU and a discrete GPU each had their own memory, and data had to be copied between them across a bus. On the M1, the CPU, GPU and Neural Engine share one pool of high-bandwidth memory in the same package. That removes copies, reduces latency and saves power. The trade-off is that memory cannot be upgraded after purchase, so buyers must choose the right amount at the start.
From a chip design point of view, the M1 showed the payoff of tight hardware and software co-design: Apple controlled the instruction set implementation, the SoC, the operating system and the compilers, and tuned them together.
What came next
Apple followed the M1 with the M1 Pro and M1 Max in October 2021 and the M1 Ultra in March 2022, then the M2 generation from June 2022, the 3 nm M3 family in October 2023 and later generations after that. The transition across the Mac range finished in June 2023 with the M2 Ultra Mac Pro.
For more details on the M1 chip, see the full M1 guide at MacRumors.
FAQ
Is the M1 an Arm processor?
Yes. It uses CPU cores that Apple designs itself under an Arm architecture licence, and it runs the 64-bit Arm instruction set.
Can M1 Macs run Intel apps?
Yes, through Rosetta 2, which translates Intel x86-64 code to Arm. Most apps run well, and native Arm versions run faster.
Who manufactures the M1?
TSMC, on its 5 nm process. Apple designs the chip; TSMC fabricates it.


Your comments will be moderated before it appears here.