Apple Weighs AI Servers Built on Its Own Chips

Apple is developing an enterprise AI server built around its own future chips and has discussed using Nvidia networking technology, Reuters reported, citing The Information. The system is reportedly aimed at running AI models, with a possible launch in 2029.

The server may use two or four M8 Ultra processors, according to the report. Those chip configurations and the launch timing are still plans, not a confirmed product; the project could change or be canceled.

Nvidia’s possible role would be connecting the processors, not necessarily supplying the main computing chips. Apple has discussed NVLink Fusion, a technology Nvidia says can link custom processors into larger AI systems.

That distinction matters for Nvidia. Apple could design the core silicon and still rely on Nvidia’s networking technology, creating a potential sale without buying Nvidia’s AI processors. No agreement has been announced, and Apple could proceed without Nvidia’s system.

The plan would extend Apple’s silicon strategy from consumer devices into equipment for businesses and AI developers. It follows reported demand from AI companies for Mac mini and Mac Studio computers, which use Apple-designed chips.

Apple’s last dedicated server line, Xserve, was discontinued in 2011. A return would put the company back in a market where buyers weigh computing performance, power use and the ability to connect machines at scale.

For Apple, selling servers could open a new customer base for its processors beyond Macs and phones. But the 2029 target leaves years for technical and commercial hurdles, and the report does not say how many systems Apple expects to sell.

Investors will be watching for signs that the project moves beyond discussions, including a confirmed product, Nvidia partnership or clearer launch schedule. Until then, the report points to a possible collaboration, not a booked source of revenue for either company.