Apple is developing an enterprise rack server powered by forthcoming M8 Ultra chips and has discussed licensing Nvidia's NVLink Fusion to connect processors at rack scale, The Information and Bloomberg reported. The project, backed by John Ternus before he became Apple's chief executive on September 1, 2026, targets a launch no earlier than 2029 but remains subject to cancellation or a redesign using Apple's own technology. The effort follows heavy AI-lab demand for Mac mini and Mac Studio systems. OpenAI bought tens of thousands of Macs for reinforcement learning and computer-use-agent training, while Anthropic has rented Mac mini capacity through Amazon Web Services. Apple's fiscal third-quarter Mac revenue rose 29% year over year to $10.4 billion for the quarter ended June 27, 2026, revising an earlier reported figure of $10.3 billion. Apple's UltraFusion packaging technology links two dies within a single chassis but is not designed for multi-server rack deployments. NVLink Fusion would provide access to Nvidia's rack-scale interconnect, though deployments require at least one Nvidia GPU or CPU. Apple also serves on the UALink Consortium board, making possible adoption of Nvidia's proprietary platform strategically notable.