In 2020 Apple began using its own desktop processors in its Macs, starting with the arrival of the M1 processor in the MacBook Air, 13-inch MacBook Pro, Mac mini, and 24-inch iMac. We’ve since seen the debut of the higher-end versions in the form of the M1 Max and M1 Pro, which arrived in the 14-inch MacBook Pro and 16-inch MacBook Pro in October 2021, and finally the M1 Ultra, which debuted with the new Mac Studio in March 2022.
Now Apple has revealed the details of the second generation of Apple silicon. We’ve gathered all the news about what to expect from Apple’s M2 system on chip, all its upcoming variants including the M2 Pro, Max and Ultra and details of the first Macs to feature them.
M2 chip family release date: When will the M2 Pro, Max and Ultra be released?
M2: July 2022
M2 Pro and M2 Max: November 2022 (expected)
M2 Ultra and M2 Extreme: 2023 (rumored)
The first M2 chips shipped with the MacBook Air and 13-inch MacBook Pro and the M2 MacBook Pro in June and July 2022.
M2 chip family specs: M2 Pro, Max and Ultra predicted specs
As with the base model of the M1, the M2 is aimed at consumers rather than professionals. Its focus is on energy efficiency as well as general performance enhancements over its predecessor. See: M2 vs M1, M1 Pro, M1 Max, M1 Ultra.
The M2 offers an 8-core CPU, 10-core GPU, as well as up to 24GB RAM. That’s more GPU cores than the M1, which had 8 GPU cores, and a higher RAM limit, which topped out at 16GB RAM on the M1.
Now that we know what is on offer from the standard M2 chip we can extrapolate some possible specs for the rest of the M2-series. These predictions are based on the same doubling up of components in the M1-series.
The M1 chips offered the following:
M1: 8-core CPU, 8-core GPU, up to 16GB RAM
M1 Pro: up to 10-core CPU, up to 16-core GPU, up to 32GB RAM
M1 Max: up to 10-core CPU, up to 32-core GPU, up to 96GB RAM
M1 UltraL: up to 24-core CPU, up to 64-core GPU, up to 64GB RAM
M1 Extreme (predicted): 48-core CPU and 128-core GPU cores, up to 128GB RAM
So based on what we know, we can assume the following specs for the M2 family of chips:
M2: 8-core CPU and 10-core GPU, up to 24GB RAM
M2 Pro: up to 10-core CPU, up to 20-core GPU, up to 48GB RAM
M2 Max: up to 10-core CPU, 40-core GPU, up to 96GB RAM
M2 Ultra: 24-core CPU, 80-core GPU, up to 192GB RAM
M2 Extreme: 48-core CPU, 160-core GPU cores, up to 384GB RAM
However, we may not see quite as high specs as the above. Mark Gurman from Bloomberg’s predicted in October 2022 that the configuration options for the M2 Ultra will include 24- and 48-core CPUs, 76- and 152-core GPUs, and up to 256GB of memory.
Based on the M2 it is possible that Apple may not increase the number of CPU cores for the rest of the M2-series, but that doesn’t necessary mean that the CPUs in the new chips will be no better than those in the M1 series. Due to changes in the way that the chips are manufactured, there are improvements seen in the CPU performance regardless of the number of CPUs.
While the manufacturing process remains at 5nm, the M2 uses the second-generation 5-nanometer technology. TSMC’s next generation N4P process, is an enhanced version of the 5 nanometer process and is supposed to deliver about 11 percentage points more performance and almost 22 percentage points more efficiency compared to the conventional 5 nanometer process (used for the production of A15 and M1, M1 Pro and Max).
Apple makes claims that seem to meet these expectations. According to Apple the M2 offers: “an 18 percent faster CPU, a 35 percent more powerful GPU, and a 40 percent faster Neural Engine,” which matches with our benchmarks. There is also 50 percent more memory bandwidth compared to M1, and up to 24GB of fast unified memory, which helps increase the overall performance.
There are reports that suggest that Apple is moving to a 3nm process, which could mean we see even more powerful variants within the M2-series due to even higher transistor density. However, this advance may not transpire until the M3 series arrives, according to Digitimes, who claimed in December 2021 that TSMC, who manufactures the chip for Apple, had to postpone its 3nm plans due to the complexities with the process. It should be noted that Apple moved to a 4nm process for the A16 in the iPhone 14, so Apple could use that in the M2 chips as well.
Based on our tests we can demonstrate the advancements made by the M2 compared to the M1, as outlined below.
Even before the M2 launched, we had seen reports that Apple would keep to the same 8 CPU cores with the M2 as it had for the M1 and this is the case, however, the improvements offered by the M2 should mean that the new Macs are much more powerful and energy efficient than the predecessor.
The N4P process, mentioned above, allows for a greater transistor density, which could allow individual cores to be clocked a little faster than in the M1. There are 25 percent more transistors than in the M1, which allow for improvements to the memory controller and memory bandwidth.
As with the M1, the M2 CPU cores are split into high performance and high efficiency variants. The faster performance cores are paired with a larger cache, while the efficiency cores see even greater performance gains, according to Apple.
Apple claims: “Compared with the latest 10-core PC laptop chip, the CPU in M2 provides nearly twice the performance at the same power level.”
Compared to PCs with more CPU cores, the M2 is able to provide almost as much performance with a fraction of the power consumption. According to Apple: “M2 provides nearly 90 percent of the peak performance of the 12-core chip while using just one-fourth the power.”
In our tests of the M2 MacBook Pro we could see that the multi-core result was still below that of the M1 Pro, although when it came to single-core it beat the M1 Pro.
Apple also expanded the graphics card performance to ten graphics cores, two more than the current M1. This was anticipated.
Apple claims the 10-core GPU delivers “up to 25 percent higher graphics performance than M1 at the same power level”, thanks to the larger cache and higher memory bandwidth. At maximum power this is 35 percent better, claims Apple.
Apple also claims that the GPU delivers 2.3x faster performance compared to the latest integrated graphics (while using a fifth of the power).
Thanks to the additional GPU cores, 10-cores rather than the 8-core limit of the M1, the M2 performs better than the M1, but it’s still below the M1 Pro with 14-cores. None of which should be a surprise.
The M2 supports more RAM than the M1 does – the M1 maxes out at 16GB RAM, while the M2 can support 24GB RAM. There are hopes therefore that the M2 Pro, Max and Ultra will support more RAM, up from 32GB, 64GB and 128GB respectively in the M1-series.
In addition, the memory controller in the M2 offers 100GB/s of unified memory bandwidth, which is 50 percent more than M1.
M2 Battery life
Apple’s claims about power efficiency might lead you to expect better battery life, but battery life for the M2-powered MacBooks is still 18 hours.
Apple still sells an Intel version of the higher-end Mac mini, so that model is rumored to finally get an update to an M2 processor this fall. Additionally, the higher-end MacBook Pro machines will likely get a spec bump at the same time.
Which Macs will get the M2 Ultra and Extreme processor?
A new Mac Studio will likely arrive with the M2 Max and M2 Ultra in 2023, but there could be a new Mac Pro in the pipeline with the M2 extreme processor. We’re not sure on the timing of this announcement, but Apple teased it release at the Peek Performance event in March.
And the M3…
With the M2 arriving about a year and a half after the M1, we can assume that the M3 chips will being rolling out in late 2023. Like the A-series iPhone chips, that processor is likely to bring a bigger jump in performance due to architectural changes and the move to a new 3nm process.
If the 3nm process isn’t used for the still to launch M2 chips it looks certain to arrive with the M3.
According to Digitimes, the first products using the 3nm processor will be released in the first quarter of 2023 with production starting at the end of 2022. Qualcomm, Samsung and Intel are all set to use the 3nm process. The Information has reported that Apple’s M3 chips have the code names Ibiza, Lobos and Palma.