Apple plots return to server market with custom AI-focused hardware

Apple is developing an enterprise server designed specifically for AI workloads, signaling the company's intention to re-enter the server market after roughly 20 years of absence. The system would leverage Apple's existing M-series Ultra architecture, the same high-performance chips currently powering Mac desktops.

Product Details#

The forthcoming server is slated for 2029 release and would be offered in two variants: one housing two M8 Ultra chips and another with four, according to reporting from The Information. The project gained momentum under support from newly appointed CEO John Ternus, who championed the initiative while overseeing Apple's hardware engineering division.

Riding a Wave of Mac Adoption#

Apple's move reflects an unexpected surge in AI developer interest in Mac hardware. Companies including OpenAI have reportedly purchased "tens of thousands" of Mac minis and Mac Studios to power reinforcement learning experiments, while Anthropic has leased similar machines through Amazon Web Services. This groundswell of enthusiasm for consumer-grade Apple computers running AI models has evidently emboldened leadership to pursue a purpose-built server offering.

Strategic Implications#

The server project represents Apple's attempt to capitalize on its inadvertent success in the AI hardware space. Rather than developing entirely new chips, the company plans to reuse its proven M-series Ultra design—a pragmatic approach that leverages existing engineering work and manufacturing expertise while meeting legitimate enterprise demand from a developer community that has already voted with its wallet.