According to The Decoder, citing The Information, Apple is working on an enterprise server for the AI inference market. Configurations under consideration include two or four M8 Ultra chips, with a potential launch not expected before 2029.
To connect the chips, Apple is considering Nvidia's NVLink Fusion technology. The Decoder cites large-scale purchases of Mac equipment by OpenAI and Anthropic for AI workloads as a potential factor supporting the project.
This is not yet an announced product: the presented material contains no statement from Apple, no technical specifications for the server, and no confirmation of timelines. Therefore, this constitutes a report on development rather than a completed launch.
editorial commentary
Why it matters
If the report is confirmed, Apple would gain the ability to compete not only with Mac supplies but also with its own server platform for AI inference. The next observable signal will be an official statement from Apple or new independent information regarding configuration and timelines. Substantial uncertainty remains due to the lack of primary confirmation and the limited nature of the available material.