Apple May Return to Server Market with Nvidia Technology

Apple is exploring a return to the server market by designing an AI-focused rack server built around its upcoming "M8" chips and by integrating Nvidia networking technology, specifically NVLink Fusion, to link multiple processors. The system would be sold to enterprise customers that want to run and serve trained AI models on-premises, with an emphasis on inference and generative response workloads. Work on the concept began about a year ago with support from John Ternus, and a notional timeline points to a possible 2029 release, though the plan is not finalized and could be canceled.

The proposal responds to rising enterprise demand for Apple silicon for AI, and to scaling challenges Apple faces with its existing Private Cloud Compute interconnects, which Apple has so far kept closed to partners. Partnering with Nvidia would give Nvidia a networking role even as Apple builds alternative processors, and it signals warmer ties after years of friction; Apple is already working to extend Private Cloud Compute to Google Cloud using Nvidia GPUs and has discussed other collaborations including Nvidia open-source models. Major obstacles remain: Apple needs stronger business support, more developer tooling and investment in its MLX framework, and a proper rackable server product line to compete effectively.