Apple is reportedly developing its first enterprise AI server since the mid-2000s, featuring two or four M8 Ultra chips with an expected release around 2029. The project has support from CEO John Ternus and could integrate Nvidia's NVLink technology for data center networking, driven by surging demand from AI companies like OpenAI and Anthropic for Apple silicon.
Background
Apple exited the server market in the mid-2000s but has seen growing adoption of its M-series chips for AI workloads among developers. The company's Mac mini and Mac Studio have become unexpectedly popular for AI training and inference tasks.
- Source
- Ars Technica
- Published
- Sep 17, 2026 at 06:02 AM
- Score
- 8.0 / 10