Apple Rumored To Be Developing Custom Silicon For AI Servers, New Chip Said To Use TSMC’s 3nm Process And Will Enter Mass Production In 2025

Omar Sohail
Apple rumored to be developing custom silicon for AI servers

Apple introduced its first custom silicon for Macs back in 2020, which was called the M1. Since then, the company has launched several variations of these chipsets that power ‘All In One’ desktops, powerful portable Macs, and workstations like the Mac Studio and Mac Pro. With the consumer hardware category, all featuring Apple Silicon, the only area where the technology giant has yet to unveil a custom solution is in AI servers. However, according to one tipster, there is now in-house silicon being developed exclusively for that.

Mass production timeline indicates that Apple’s custom AI server chip will use TSMC’s 3nm ‘N3E’ process

A Weibo user whose account handle is machine translated to ‘Mobile chip expert’ claims that Apple is developing a custom chip intended to power AI servers. Additional specifications of this silicon are not mentioned, but he does state that TSMC’s 3nm process will be used to design the chip. Bear in mind that the Taiwanese semiconductor giant has several iterations of its 3nm technology, with the ‘N3B’ variant being the first one that Apple used to mass produce the A17 Pro last year.

Related Story Latest iPhone 16 Molds Show A Redesigned MagSafe Charging Ring, But Existing Accessories Will Likely Be Compatible

Given that the tipster highlights that this custom chip will undergo mass production in the second half of 2025, it is likely that TSMC’s 3nm ‘N3E’ node will used. The same technology is said to fabricate the A18 Pro, the upcoming Snapdragon 8 Gen 4, and Dimensity 9400, so it should not be surprising to learn that the AI server chip will be made on the same lithography. As for what Apple’s end goal is in developing a custom silicon for this category, those details are currently in the dark.

However, we have reported that Apple may bring on-device processing for AI-related tasks in iOS 18 instead of relying on the cloud for the same function. Devices that can handle on-device processing will also be equipped with cutting-edge chipsets, such as the A18 Pro, with potentially larger Neural Engines to make this possible. As for the older SoCs that will lack sufficient processing power to support on-device AI functions, cloud-based processing will be necessary. Apple’s AI server chips may improve the performance of speedy cloud-based processing and other AI-centric features.

Unfortunately, the tipster provided little to no details, making it pointless to speculate about Apple’s actual intentions with its in-house AI server chip. We recommend treating this rumor with a pinch of salt, and we will get back to you with more updates.

News Source: Mobile chip expert

Share this story

Deal of the Day

Comments