Apple developing enterprise AI server with M8 Ultra chips for 2029 launch

Sep 16, 2026

Apple is reportedly developing an enterprise server powered by its future M8 Ultra chips, aiming for a launch in 2029. The project, backed by CEO John Ternus, marks Apple's potential return to the server market since retiring its Xserve line in 2011. Designed for AI inference, the server will offer configurations with two or four M8 Ultra chips and may utilize Nvidia's NVLink Fusion technology. This development follows surging demand from AI firms like OpenAI and Anthropic, which are buying Mac hardware in bulk.

Apple M8 Ultra server

▪Apple Inc. is developing an enterprise server using its own chips, designed for artificial intelligence inference to run already-trained models, with a planned launch not before 2029
▪Apple's planned enterprise AI server, targeted for a 2029 launch, will be offered in two configurations, featuring either two or four of Apple's future M8 Ultra chips
▪Apple CEO John Ternus supported Apple's enterprise AI server project targeted for a 2029 launch approximately a year before September 2026 while he was still serving as Apple's head of hardware engineering

AI compute hardware demand

▪The growing demand for compute power in the expanding artificial intelligence industry has led Apple to target AI developers, businesses, and governments with its ARM-based M processors
▪Apple's Mac revenue increased by nearly 29 percent in the quarter preceding September 2026, reaching $10.4 billion

Enterprise server market re-entry

▪The expected 2029 release of Apple's planned enterprise AI server, powered by M8 Ultra chips, would mark Apple's first server product to hit the market in nearly two decades
▪Apple retired its Xserve server line in 2011 and has largely left the enterprise machine market to other manufacturers since then

Mac Mini AI workloads

▪Anthropic has rented Mac Minis from Amazon Web Services to support its artificial intelligence workloads
▪Artificial intelligence labs, including OpenAI and Anthropic, are purchasing or renting Apple Mac Minis and Mac Studios in bulk to run AI workloads
▪OpenAI has purchased tens of thousands of Mac Minis and Mac Studios to train artificial intelligence agents through trial-and-error reinforcement learning

Debatable claims

▪Apple's M-series chips will challenge Nvidia's dominance in the AI hardware market
▪Consumer-grade hardware is a viable alternative to dedicated enterprise servers for AI development