Apple Plans M8 Ultra Server for AI Workloads by 2029
The company's first enterprise server in nearly two decades would target AI developers already buying Mac hardware in bulk.
Apple is developing a dedicated AI server built around its M-series Ultra chips, with a planned 2029 release that would represent the company's first enterprise server since the Xserve was discontinued in 2011.
The server would ship in two configurations featuring either two or four M8 Ultra chips, according to details first reported by The Information. The project received backing from John Ternus—who became Apple CEO on September 1, 2026, succeeding Tim Cook—when development began a year ago during his tenure leading hardware engineering.
Strong Mac sales to AI developers drive the push
The server initiative builds on unexpected momentum in Apple's desktop lineup. AI companies have been purchasing Mac minis and Mac Studios in substantial volumes to run AI workloads, particularly for training AI agents through reinforcement learning methods.
The Information reports that OpenAI has acquired tens of thousands of Mac minis and Mac Studios for this purpose, while Anthropic has rented Mac minis through Amazon Web Services. This enterprise adoption of consumer Mac hardware has validated demand for Apple silicon in AI infrastructure.
Nvidia partnership under consideration
Apple is exploring the use of Nvidia's data center networking equipment to connect the M8 chips in its planned servers. Sources indicate the companies have discussed how Apple might leverage Nvidia's NVLink Fusion technology for its AI initiatives.
The Information notes the project could proceed without Nvidia's involvement or be canceled entirely, but the conversations signal Apple's willingness to integrate third-party technologies where strategic.
Supply chain constraints loom
The server plans face the same memory chip shortage affecting the broader tech industry. Companies racing to build AI data centers have created sustained demand that has driven up prices across consumer electronics, including gaming devices, flagship smartphones, and Apple's existing product line.
Why it matters
Apple's potential return to enterprise servers reflects a fundamental shift in how AI infrastructure is being built. The strong sales of Mac hardware to AI companies demonstrate that developers value Apple silicon's performance and efficiency characteristics for specific AI workloads, particularly reinforcement learning. If Apple delivers a purpose-built server, it could establish a new revenue stream in the enterprise market while providing AI companies with hardware optimized for their needs—though success depends on navigating memory supply constraints and competing against established data center vendors.
The details were first reported by The Information.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call
