Apple developing enterprise AI servers on its own M-series Ultra chips, first dedicated rack servers since Xserve, targeting developers, businesses and governments as early as 2029

Apple is developing enterprise AI servers powered by its own M-series Ultra chips, its first dedicated rack servers since it discontinued Xserve in 2011, aimed at companies and governments that want to run AI models on their own machines rather than public clouds, and potentially launching as early as 2029, according to reports. Systems would run two or four Ultra chips for private AI inference. Mark Gurman, who first reported the plan, said the 2029 server could be built around the M7 Ultra rather than the M8 Ultra unless Apple is developing both simultaneously, and that a separate Apple AI server is due to go live in early 2027. The move follows Mac revenue rising nearly 29% last quarter to $10.4B, driven by AI developer demand for Mac mini and Mac Studio, and would expand Apple's silicon strategy beyond consumer devices into the enterprise infrastructure market, challenging Nvidia in the data center for the first time. Apple shares rose on the report.

Detected & updated continuously · Source: Nebula