Apple is reportedly developing an enterprise AI server using its M-series Ultra chips, with a potential release in 2029. The server would feature either two or four future M8 Ultra chips and mark Apple’s first server launch in nearly two decades. The project received early support from John Ternus, who led hardware engineering before becoming CEO.

This move aligns with growing demand from AI developers for Apple’s Mac mini and Mac Studio systems, which are being used to run AI workloads. Companies like OpenAI and Anthropic have purchased or rented tens of thousands of these Macs for training AI models via reinforcement learning. The trend has encouraged Apple to create a dedicated server solution leveraging its high-performance chip architecture.