Reports from The Information and other sources have revealed that Apple is advancing the development of dedicated AI servers utilizing M8 Ultra chips.
Apple reportedly considering development of AI servers using NVIDIA technology, aiming for release around 2029
This article is a translation. Read the Japanese original
Under the company's plan, NVIDIA's NVLink Fusion is expected to be adopted. NVLink Fusion is a platform designed to directly connect third-party proprietary CPUs or custom AI accelerators to NVIDIA's NVLink. By supporting this technology, Apple will be able to connect Apple Silicon to rack-scale AI solutions.
Apple is currently reportedly developing 2 types of servers; one configuration is expected to feature 1 M8 Ultra chips with 2 units, while another will consist of 1 units with 4 units. Product availability is anticipated around 2029.
If realized, this would mark Apple's re-entry into the server market since the discontinuation of the Xserve in 2011. On the other hand, Todd Daily, who has experience working at Apple, expressed skepticism regarding the move, citing the challenges of supporting legacy operating systems and the lack of enterprise-grade maintenance infrastructure as reasons why the company previously abandoned its server business.
Sources
- AppleはNVIDIAのNVLink Fusionを介して接続されたM8 Ultraチップを使用したAIサーバーを開発しており外部に販売することも計画中 （GIGAZINE、2026-09-17）