Apple plans AI servers with M8 ultra chips and Nvidia technology: What to know

Apple is reportedly building a business AI server around its own M8 Ultra chips targeting a 2029 launch and is currently in talks with Nvidia about using the chipmaker’s NVLink Fusion integrated platform to link those chips together, according to a report by The Information published on Wednesday.

Nvidia, which commands the AI server market and pulls in roughly a fifth of its data-centre revenue from networking gear, is the primary beneficiary of that development. Notably, Apple's potential deal to adopt NVLink fusion would secure an anchor account while simultaneously validating Nvidia’s key strategy of monetizing interconnect even alongside its dominant GPU franchise.

According to reports, the planned server would come in two configurations with one pairing two M8 Ultra chips and one pairing four and focuses on AI engineers, businesses and governments. Some Apple engineers developed a project involving Nvidia’s NVLink Fusion, a technology used by software users to allow chips from different vendors to communicate at high speeds.

The project is driven by a clear commercial necessity with Mac being Apple’s fastest-growing product segment in its most recently recorded quarter where revenue soaring nearly 29% to $10.4 billion sparked by an unusual demand from AI developers buying up Mac mini and Mac Studio units.