Apple Reportedly Planning Enterprise AI Servers Powered by M8 Ultra Chips by 2029

Apple is reportedly planning to re-enter the enterprise server market around 2029 with custom AI servers featuring up to four unreleased M8 Ultra chips. The project may also leverage Nvidia's high-speed NVLink Fusion interconnect technology to allow multiple Apple Silicon processors to work together seamlessly. This initiative signals Apple's ambition to capture the growing market for on-premise AI inference hardware among enterprises and research labs. Furthermore, adopting NVLink highlights a significant thawing in Apple's historically strained relationship with Nvidia as interconnect technology becomes critical for AI infrastructure. The planned server line includes a base configuration equipped with two M8 Ultra chips and a high-end variant housing four chips, aimed primarily at AI inference workloads rather than cloud training. However, because the target launch date is distant (around 2029), the project remains subject to potential alterations or cancellation.