Apple is reportedly developing an enterprise-grade AI server powered by its own chips and is considering using Nvidia’s networking and chip-interconnect technology, according to people familiar with the project.

The server would be marketed to AI developers, businesses and government agencies. Apple is reportedly planning two versions: a lower-end model with two M8 Ultra chips and a higher-end model with four. The M8 Ultra remains under development and would belong to Apple’s most powerful chip family, currently used in products such as the Mac Studio.

Apple is also considering Nvidia’s complete NVLink Fusion platform, including switches, chiplets and supporting software. The technology would enable multiple chips inside the server to exchange data at high speeds. Some Apple employees involved in the project reportedly view Nvidia’s system as the industry’s leading chip-interconnect technology.

Under the current plan, the server could reach the market in 2029. However, the project remains at an early stage and could still be canceled or launched without Nvidia technology.

Apple’s new CEO, John Ternus, reportedly supported the project when it began about a year ago, while he was still leading the company’s hardware engineering division.