Apple reportedly developing AI-focused enterprise server with M8 ultra chips
Key Points
- Apple is reportedly developing an AI enterprise server using its own M8 Ultra chips, for sale to AI developers, businesses, and governments.
- Apple has discussed using Nvidia's NVLink Fusion networking tech, though this and the project itself remain unconfirmed and could be canceled.
- The server would be Apple's first since Xserve ended in 2011, with a possible 2029 launch.
Apple is reportedly developing an enterprise server built around its own chips for sale to AI developers, businesses, and governments, according to a report from The Information cited by Bloomberg and Reuters. The server would use Apple's forthcoming M8 Ultra chips and could arrive in two configurations, one with two M8 Ultra chips and a higher-performance version with four. Apple has also reportedly discussed using Nvidia's networking technology to connect the chips, though this involvement remains under discussion rather than confirmed. Apple and Nvidia did not respond to requests for comment, and Reuters said it could not independently verify the report.
What materially changed: Apple's return to server hardware
If it proceeds, this would be Apple's first dedicated server product since it discontinued its Xserve line in 2011, when the company shifted its hardware focus entirely to consumer devices and standard Mac hardware. The project has reportedly been underway for about a year and was backed from the start by John Ternus, then Apple's hardware chief and now its CEO. The Information reported the server likely would not reach the market until 2029, and the effort could still be canceled or launched without Nvidia's technology.
Essential context: Nvidia's networking role and a thawing relationship
The chip connection under consideration is Nvidia's NVLink Fusion, technology that gives chipmakers pre-built connectors and specialized memory so custom AI chips can link into Nvidia's larger data-center systems. Some Apple staff reportedly believe Nvidia offers the best available connectivity option for the project. Apple and Nvidia have had a strained relationship for close to two decades, dating back to disputes over MacBook graphics chip failures that led Apple to sever ties with the company. A report of renewed technical discussions would mark a shift in that relationship. Apple refreshed its Mac lineup with M6, M5 Pro, M5 Max, and M5 Ultra chips last month, with the company saying its 2-nanometer M6 chip delivers up to four times faster AI performance than its predecessor.
Why it matters for enterprise AI buyers
A dedicated Apple AI server would put the company in direct competition for enterprise inference workloads, the growing category of running trained AI models in production rather than training them, as businesses shift spending from model development to deployment. It would also extend Apple's in-house silicon strategy beyond consumer devices and into infrastructure sold directly to AI developers and governments. The move comes as capital continues flowing into AI infrastructure and compute capacity more broadly, a trend also visible in Firmus Technologies' plan to raise up to $5 billion in an Australian IPO to fund its AI data center buildout. Any Apple entry into this market remains speculative at this stage, based on a single report citing unnamed sources.
What happens next
The Information reported the server is not expected to launch before 2029, and its final design, including whether it ultimately uses Nvidia's technology, has not been finalized. The project could still be shelved entirely. Apple has not commented on the report.