Updated
Updated · Ars Technica · Sep 16
Apple Develops 2029 AI Server With Up to 4 M8 Ultra Chips
Updated
Updated · Ars Technica · Sep 16

Apple Develops 2029 AI Server With Up to 4 M8 Ultra Chips

3 articles · Updated · Ars Technica · Sep 16

Summary

  • Apple is building an enterprise AI server for a possible 2029 launch, its first server product in nearly 20 years.
  • The machine would come in two configurations using either two or four future M8 Ultra chips, extending the high-end M-series processors now used in Mac desktops.
  • John Ternus backed the project when it began about a year ago, before he became Apple’s new CEO, according to The Information.
  • Demand from AI developers appears to be driving the push: OpenAI has bought tens of thousands of Mac minis and Mac Studios for reinforcement-learning workloads, while Anthropic has rented Mac minis from AWS.
  • The effort would mark Apple’s return to a market it exited after retiring the Xserve line in 2011, as AI computing demand reshapes enterprise hardware.

Insights

Why would Apple partner with long-time rival Nvidia to build massive AI servers after abandoning the market 15 years ago?
As data sovereignty demands rise, will Apple's highly secretive M8 Ultra servers become the ultimate fortress for private enterprise AI?