WWDC 2022

Apple's M2 chip explained in 5 key stats

The next generation of Apple's M-series chips has arrived. Here's everything you need to know.

Apple

Since Apple’s M1 chips were unveiled in 2020, it hasn’t wasted any time rolling out upgrades. Last year we got the M1 Pro and the M1 Max, and recently the M1 Ultra. At WWDC 2022, Apple one-upped itself with its next-gen M2.

SOPA Images/LightRocket/Getty Images

Here’s the M2 explained in five key stats...

20 B

The M2 crams 20 billion transistors into one chip.

wera Rodsawang/Moment/Getty Images

With 25 percent more transistors than the M1, the M2 will get a significant performance boost. If 20 billion transistors on a chip doesn’t impress you, just consider for a moment that those transistors are somehow crammed onto Apple’s 5nm (yes, nanometers) architecture. That’s a lot of heat for not a lot of wafer.

18%

The performance boost of the M2 over the M1.

Naturally, with that many transistors, the M2 will get a significant performance boost. Apple says the M2 will have 18 percent greater performance over the M1. The M2 also comes with Apple’s new GPU, which has 10 cores — AKA two more than M1.

SOPA Images/LightRocket/Getty Images

35%

Greater graphics performance compared to the M1.

Speaking of GPUs, Apple says that the M2 delivers up to 35 percent greater graphics performance at its max power. Like the M1, the M2 delivers while consuming less power. Apple says that compared to the integrated graphics on the “latest PC laptop chip” its M2 GPU brings 2.3x faster performance at the same power level.

1/4th

The fraction of power the M2 uses compared to *other* 12-core chips.

gece33/iStock Unreleased/Getty Images

The M2 continues the tradition of lapping up rival chipmakers on energy-to-performance. According to Apple, the M2 delivers about 90 percent of the peak performance of a leading 12-core chip (looking at you, Intel) while only using a quarter of the power. That means speed without totally butchering battery life.

15.8 T

The M2 performs 15.8 trillion operations per second.

The M2 can process an impressive 15.8 trillion operations per second, making its neural engine 40 percent faster than the M1. That’s great for anyone using an M2 device for things like video or image processing — or really anything involving machine learning for that matter.

seksan Mongkhonkhamsao/Moment/Getty Images