Apple Develops Enterprise AI Server With M8 Ultra Chips, Reportedly Relying on Nvidia Networking

Apple (AAPL) is developing an enterprise AI server built on its own M8 Ultra chips, according to The Information, marking its first return to the server market since discontinuing Xserve in 2011. The machines would target AI developers, enterprises and governments seeking on-premises inference hardware, with launch not expected until 2029. The plan reportedly depends on Nvidia (NVDA) networking technology to link chips, as Apple's internal chip-to-chip connection is said to be too slow and costly at server scale. The shift softens Apple's prior on-device AI stance and follows Mac segment revenue rising 29% year-over-year to $10.4 billion last quarter, driven partly by AI labs buying Mac minis and Studios in bulk. Apple reported fiscal third-quarter revenue of $109.42 billion, up 16% year-over-year, with EPS of $2.20. It guided fourth-quarter revenue growth of 9% to 11%. AAPL trades near $322, up 40% over the past year, with a forward P/E of 37.59x.