Apple Eyes Custom M8 AI Servers via NVLink Tie-Up

Written by Sandra Leggero

Sandra Leggero has a background in financial markets, having spent more than 9 years in commodities trading for several European and Asian companies. She holds a degree in Economics from the University of Pavia and specializes in emerging markets.

, | Updated: September 16, 2026

Key Moments

Apple is working on an enterprise AI server built around in-house M8 Ultra chips, targeting a 2029 launch, according to The Information.
The company has reportedly held talks with Nvidia (NASDAQ: NVDA) about using NVLink Fusion to connect M8 Ultra chips in dual- and quad-chip server configurations.
Surging demand from AI developers pushed Mac revenue up nearly 29% to $10.4 billion in Apple’s most recently reported quarter, highlighting constraints of consumer hardware for enterprise AI.

Apple’s Long-Term AI Server Ambitions

Apple is reportedly building an enterprise-focused AI server platform powered by its own M8 Ultra processors, with an internal goal of bringing the product to market in 2029, The Information reported Wednesday.

The envisioned system is designed for high-end artificial intelligence workloads in corporate, developer, and government environments. It marks a move by Apple to translate its chip design capabilities into dedicated AI server infrastructure, rather than relying solely on consumer-grade Macs for AI development activity.

Potential Nvidia NVLink Fusion Partnership

As part of the effort, Apple has held discussions with Nvidia (NASDAQ: NVDA) about integrating Nvidia’s NVLink Fusion interconnect technology to tie multiple M8 Ultra chips together, The Information reported.

NVLink Fusion includes switches, chiplets, and software engineered to enable high-speed communication between processors from different vendors. Some Apple engineers involved in the project reportedly view NVLink Fusion as the strongest available connectivity option for linking the M8 Ultra chips in the server platform.

Nvidia currently dominates the AI server landscape and generates roughly one-fifth of its data-center revenue from networking equipment, according to the report. A decision by Apple to adopt NVLink Fusion would give Nvidia a significant new volume customer while reinforcing its approach of monetizing interconnect technologies even as competing processors emerge in the market.

Planned System Configurations

The Information reported that Apple’s planned AI server would be offered in two main configurations:

Configuration	Processor Setup	Target Users
Option 1	Two M8 Ultra chips	AI developers, businesses, governments
Option 2	Four M8 Ultra chips	AI developers, businesses, governments

The aim is to create a scalable platform for AI-intensive workloads by linking multiple M8 Ultra processors with advanced interconnect technology. However, The Information noted that the initiative could still be cancelled, and neither Apple nor Nvidia has publicly acknowledged the reported talks.

Leadership Backing From New CEO John Ternus

New Apple (NASDAQ: AAPL) CEO John Ternus, who assumed the role earlier this month, has been closely associated with the AI server initiative since its early days, The Information reported. While leading Apple’s hardware engineering organization roughly a year ago, Ternus supported the early development of the project, according to people familiar with the matter cited by The Information.

His ongoing advocacy inside the company is seen as a key factor in sustaining momentum for the effort, suggesting internal alignment at senior levels around Apple’s longer-term enterprise AI hardware strategy.

AI Demand Drives Mac Growth and Exposes Constraints

The Information reported that the AI server project is rooted in a clear commercial need. In Apple’s most recently reported quarter, Mac was the company’s fastest-growing product category, with revenue jumping nearly 29% to $10.4 billion. That growth was driven in part by stronger-than-expected demand from AI developers purchasing Mac mini and Mac Studio systems.

This surge in AI-oriented usage has led to notable supply tightness and highlighted the limitations of consumer-grade machines for enterprise-scale AI workloads. The reported M8 Ultra-based server initiative is positioned as a response to these constraints, aiming to provide purpose-built infrastructure for high-end AI applications.