Apple to develop custom chips for AI servers

In 2020, Apple unveiled its first custom chip for Mac, the M1, heralding a new era of tailor-made silicon for various Mac models. By June 2023, this initiative culminated in the release of a new Mac Pro equipped with the M2 Ultra chip, completely phasing out Intel’s x86 processors from its product line.

According to a report by Wccftech, Apple is developing a custom chip designed to power artificial intelligence (AI) servers. The specific specifications and objectives for this chip remain undisclosed. It is rumored that Apple has chosen TSMC’s 3nm process node for manufacturing this chip, with mass production expected in the latter half of 2025. If the timeline and TSMC’s semiconductor process advancements align, this could correspond to the N3E process.

Two months ago, reports surfaced that Apple had officially abandoned its decade-long, massively funded “Project Titan” electric vehicle initiative. Subsequently, Apple disbanded its development team of approximately 2,000 members, reallocating them to other areas, notably the artificial intelligence division. Rumors suggest that Apple has shifted its focus to generative AI, aiming to discover new avenues for growth within its operations.

Rumors also indicate Apple’s plans to invest heavily in integrating artificial intelligence across its entire product line, with a particular emphasis on providing AI hardware acceleration within the Mac product family. The primary objective of the forthcoming Apple M4 series is to enhance the processing capabilities for AI applications, accelerating its research and development efforts. Moreover, this year’s iOS 18 update is expected to introduce device-end processing related to artificial intelligence, with further integrations and optimizations in software anticipated in the future.