Apple’s Chief Operating Officer, Jeff Williams, reportedly embarked on a clandestine journey to Taiwan, where he held a secretive meeting with C.C. Wei, the CEO of Taiwan Semiconductor Manufacturing Company (TSMC). The purpose of this covert rendezvous was purportedly to discuss Apple’s interest in reserving the first batch of TSMC’s highly anticipated 2-nanometer production capacity. Such a move underscores Apple’s unwavering commitment to staying at the forefront of semiconductor technology, crucial for enhancing the performance of its M-series chips and maintaining competitiveness in the rapidly evolving landscape of artificial intelligence (AI).
Sources from the Taiwanese news outlet Economic Daily divulged insights into these covert discussions, albeit under the condition of anonymity, reflecting Apple’s characteristic discretion when it comes to supplier negotiations. Indeed, Apple has long maintained a veil of secrecy around its dealings with suppliers, rarely divulging details of such strategic engagements to the public.
The significance of this meeting is further underscored by Apple’s recent foray into the AI server domain. The Cupertino-based tech giant has been actively developing proprietary AI chips tailored for its data center servers, essential infrastructure for handling AI processing tasks too complex for on-device execution, as typically encountered in iPhones or iPads. Speculation abounds that the forthcoming iOS 18, slated for release later this year, will heavily leverage advanced AI models, mirroring Android’s adoption of Gemini Nano technology.
Confirming Apple’s commitment to AI-driven innovation, the company’s Chief Financial Officer, Luca Maestri, disclosed Apple’s ownership of dedicated data centers geared towards AI applications. While Apple also leverages third-party data centers, Maestri emphasized the company’s ongoing efforts to bolster its in-house data center infrastructure. He further revealed that Apple has allocated a staggering sum of over $100 billion towards generative AI research and development initiatives over the past five years, underscoring the company’s steadfast commitment to AI innovation.
Presently, Apple relies on TSMC’s cutting-edge 3-nanometer process for its M4 System-on-a-Chip (SoC), integrating a more robust Neural Processing Unit compared to its predecessors. To further augment its AI capabilities, Apple purportedly plans to develop three additional iterations of the M4 SoC, colloquially known as Donan, Brava, and Hydra, each boasting enhanced AI processing power.
Anticipated for mass production in the latter half of 2024, these new iterations of the M4 chip represent a significant leap forward for Apple’s AI ambitions. Furthermore, Apple’s prospective acquisition of TSMC’s inaugural 2-nanometer production capacity could potentially translate into a substantial revenue boost for TSMC, with estimates pegging it at an all-time high of NT$600 billion (approximately USD $18.6 billion). Over time, Apple’s continued collaboration with TSMC holds the potential to propel the chipmaker’s annual revenue to NT$1 trillion (USD $31 billion), reflecting the magnitude of this strategic partnership.
However, both Apple and TSMC have maintained a tight-lipped stance regarding the nature of Jeff Williams’ visit to Taiwan, refraining from commenting on the specifics of their discussions. This silence underscores the confidentiality surrounding strategic deliberations between two industry giants, leaving analysts and industry observers to speculate on the potential implications of this enigmatic rendezvous.