Apple is gearing up to introduce AI inference servers that will be outfitted with its self-developed chips, signaling a potential re-entry into the server market. These servers are predominantly targeted at AI developers, corporate clients, and government entities. They are anticipated to incorporate two or four M8 Ultra chips, which are currently in the development phase. The inaugural AI server chip from Apple bears the internal codename "Baltra," a product of its collaboration with Broadcom. It leverages TSMC's cutting-edge 3-nanometer N3E process and Chiplet architecture for enhanced performance.

Furthermore, Apple is contemplating the integration of NVIDIA's NVLink Fusion networking technology, slated for a 2025 release. This move aims to facilitate high-speed interconnection between M8 chips and NVIDIA GPUs within the servers, thereby optimizing data transfer and processing speeds.

Apple's previous foray into the server market was with the Xserve, launched in 2002. However, due to a lack of adequate emphasis on the enterprise sector, the product was discontinued in 2011. Since then, Apple has redirected its business strategy towards the consumer electronics domain. Lately, the Mac business has experienced robust growth, propelled by substantial purchases from AI firms. This trend has emerged as a pivotal factor driving Apple's foray into the enterprise-level AI server market.

Following the announcement, Apple's stock price initially experienced a slight decline but swiftly rebounded. It turned positive in pre-market trading and ultimately achieved a marginal increase, reflecting investor confidence in Apple's strategic move.