iPhones, iPads, and Macs seem set for even steeper performance and power management gains come 2022, as it looks like the company’s plans to switch to 3-nanometer (nm) chips are falling into place.
Apple’s processor development road map
We’ve considered the road ahead for Apple Silicon, and those predictions seem to be on the right track for a late 2022 shift to chips manufactured using 3nm processes. Chips manufactured with this process deliver 30% improvements in power consumption and 15% better performance in comparison with 5nm chips.
Apple designs and develops processors in house and hires chip supplier TSMC to manufacture the majority of those it uses in its devices. TSMC is investing in 3nm processor production, and a report from DigiTimes tells us the company is on track to begin early-risk production of chips based on that process in the second half of the year, with production of its existing 5nm chips expected to increase quite dramatically in the second half of 2021.
(The latter makes sense, given the majority of Macs and all iPhones and iPads Apple sells will use those chips by that time.)
M-powerment and the Mac
Macs using M1 chips currently have eight cores, with that number expected to climb to as many as 16 cores in the M1X and 32 cores in high-end Macs by 2022. Is it possible those 32-core chips will be manufactured using 3nm process technology?
Apple has promised to complete its transition to Apple Silicon in Macs by around 2022.
Most enterprise IT purchasers should note that the emerging road map suggests high-end Macs may be on track to outclass high-end PCs in terms of performance, even as Apple continues to consume a greater slice of global component and processor production.
The effect of growing global demand for processors was most recently made visible in the car industry, when Volkswagen, Ford, Fiat Chrysler, Subaru, General Motors, Honda, Renault, and Toyota were all forced to reduce production of new vehicles due to short chip supplies. (It is unclear whether demand for components used in Apple’s products contributed to the problem.)
Control of the means of production
It’s important to control the means of production, and while Apple doesn’t precisely own the factories, it is a huge client for those who do. When it comes to chips, Apple is TSMC’s biggest client — taking most of the chips the company manufactures. And we’ve previously learned it has already booked production of 3nm chips. This puts wind behind the sails of Apple’s migration to Apple Silicon, and means Macs, iPhones, and iPads will continue to lead the industry in terms of performance per watt.
Apple’s previous Mac processor supplier, Intel, is not expected to introduce chips based on a 3nm process before 2025.
The M1 chips are already impressive and the M1X processors are expected to build on this. But Macs and other devices powered by chips built on 3nm process tech will deliver even more compelling power and performance advantages as they appear, likely beginning with iPhone.
The road ahead
With a view to other aspects of Apple product development, the iPhone maker is working with TSMC on ultra-advanced display technology using micro OLED displays for use in future devices, Nikkei Asia has claimed.
The two partners are also working on micro LED tech, with trial production lines for both already in place. That work is allegedly being led by veterans hired from AU Optoelectronics. Apple is also developing its own 5G radio chips for use in future devices, with these also expected to debut around 2022-2023, and continues to finesse its GPU technologies for Macs.
What comes next? We know the industry is exploring 2nm and 1.5nm production processes. But the cost of that work means many are trying what Apple already does, striving to develop advanced SoC processors, working with new transistors, materials, and lithographic processes at an increasingly tiny, molecular, and sub-molecular scale.
In other words, another tangent to Apple’s processor development future will arrive in the form of advanced microprocessor architectures. That means we’ll be hearing keynote buzzwords and statistics such as transistor counts and operations per second designed to show how Apple’s devices deliver advantages at a molecular and sub-molecular level.
After all, at some point there will be little advantage to be gained from smaller die sizes, which is when focus will turn to further optimizations at that scale. Apple appears to be in the cat bird seat from which its next decade of processor production and development seems clear across all its platforms.
And that’s even before considering the kind of processor power, speed, and built-in neural engines we’ll need to safely handle the millions of data points required for safe and autonomous control of any Apple Car.