Apple reportedly planning enterprise AI servers with M8 Ultra chips and Nvidia NVLink Fusion, targeting 2029
Per The Information, Apple is weighing a return to the server market with enterprise AI inference servers built on two or four M8 Ultra chips, possibly linked via Nvidia's NVLink Fusion, debuting no earlier than 2029.
According to The Information, Apple is considering re-entering the server market it exited in 2011 with the Xserve retirement, developing an enterprise AI inference server aimed at developers, businesses, and governments. The server is planned in configurations with two or four future M8 Ultra chips and may incorporate Nvidia's NVLink Fusion interconnect; Apple recently turned to Nvidia chips for its revamped Siri servers. The project reportedly began about a year ago and is backed by new CEO John Ternus, who supported the effort while leading hardware. Market signals motivated the move: Mac Minis and Mac Studios are already popular with AI developers, causing shortages — OpenAI reportedly bought tens of thousands of Mac minis and Studios for reinforcement learning, and Anthropic has rented Mac minis through Amazon Web Services. Apple's Mac revenue also rose nearly 29 percent to $10.4 billion last quarter. A product would likely debut around 2029, though the plans are rumored, unconfirmed, and could be changed or scrapped. Note a framing discrepancy: Ars Technica describes it as Apple's first server offering in nearly 20 years, while The Verge and The Decoder date the exit to the 2011 Xserve retirement (roughly 15 years before 2026).
- All three reports cite The Information (published 2026-09-16 by The Verge, The Decoder, and Ars Technica)
- Enterprise AI inference server targeted at developers, businesses, and governments
- Planned configurations use two or four future M8 Ultra chips
- Nvidia's NVLink Fusion interconnect is under consideration
- Launch would come no earlier than 2029; the rumored project could still be changed or scrapped
- Project reportedly began about a year ago with backing from John Ternus, then hardware chief and now CEO
- Apple exited the server market in 2011 with the Xserve retirement; Ars Technica frames the new product as its first server in nearly 20 years
- OpenAI reportedly bought tens of thousands of Mac minis and Mac Studios for reinforcement learning
Coverage timelineoldest first · each row is one article
- · 7h agoApple might make servers again to cash in on the AI rush
The Verge · AI· 35
Apple reportedly plans AI servers pairing M8 Ultra chips with Nvidia NVLink Fusion, targeting a 2029 debut.
- · 6h agoApple is reportedly building an enterprise AI server with its own M8 Ultra chips
The Decoder· 35
Apple reportedly develops an enterprise AI inference server with two or four M8 Ultra chips, possibly using Nvidia NVLink Fusion, launching no earlier than 2029.
- · 3h agoApple reportedly building server packed with M-series Ultra chips for AI
Ars Technica · AI· 50
Apple is reportedly developing an enterprise AI server with two or four future M8 Ultra chips, targeting a 2029 release.