Apple is preparing an enterprise server with M8 Ultra chips for AI inference, potentially in cooperation with Nvidia, with a launch in 2029 at the earliest
According to The Information, Apple is developing an enterprise server for AI inference with two or four M8 Ultra chips, potentially with NVLink Fusion technology from Nvidia. The project was backed by current CEO John Ternus, and a launch is not planned before 2029.
According to The Information, the server is expected to come in two versions, with two or four M8 Ultra chips, and is aimed at AI inference, meaning running already trained models, not training them. According to the same source, John Ternus, the current CEO of Apple, backed the project about a year ago, when he was still leading the hardware division. According to The Decoder, revenue from Mac computers at Apple rose by roughly 29 percent to 10.4 billion dollars in the last quarter, which appears to have given the project greater urgency.
Apple is considering NVLink Fusion technology from Nvidia to connect the chips inside the server. The technology was originally intended for Nvidia chips but is now also open to third-party hardware. According to the available information, the server would not reach the market before 2029; the project could still be canceled or completed without Nvidia technology. Apple only introduced the M6 series chips in August 2026, so the M8 series is still further in the future. The Verge adds that relations between Apple and Nvidia have historically been cool, but Apple recently used Nvidia chips for revamped Siri servers, which could foreshadow further cooperation.
It would be the first server from Apple in nearly twenty years since the Xserve line was discontinued in 2011. According to The Verge, Mac mini and Mac Studio computers have become so popular among AI developers that this has led to shortages on the market. Ars Technica adds that OpenAI bought “tens of thousands” of Mac mini and Mac Studio computers to train AI agents using reinforcement learning, while Anthropic rents Mac mini computers through Amazon Web Services.
Why it matters
If the project reaches the product stage, it will offer AI companies an alternative to GPU servers from Nvidia and AMD based on Apple Silicon chips – according to sources, companies such as OpenAI and Anthropic are already competing for Mac mini and Mac Studio computers for AI tasks. Potential cooperation with Nvidia on NVLink Fusion could also mean closer ties between two companies that have mostly competed until now. None of this is certain, however – the release is planned for 2029 at the earliest, and the project could be canceled.
What was added since the original report
Verified updates
-
The server exists in two versions: with 2 or 4 M8 Ultra processors; Focused on AI inference (running trained models, not training them); John Ternus (the new CEO of Apple) backed the project about a year ago while leading the hardware division; OpenAI and Anthropic are already buying Mac Mini and Mac Studio for AI workloads, signaling demand; The project may have recently accelerated
- The server exists in two versions: with 2 or 4 M8 Ultra processors
- Focused on AI inference (running trained models, not training them)
- John Ternus (the new CEO of Apple) backed the project about a year ago while leading the hardware division
- OpenAI and Anthropic are already buying Mac Mini and Mac Studio for AI workloads, signaling demand
- The project may have recently accelerated
Relevant practical impact
What this means
For a business
Companies planning their AI infrastructure for the long term can monitor whether Apple actually enters the enterprise AI server segment – from 2029, this could provide an alternative to current GPU servers from Nvidia and AMD, but the project could still be canceled.
StrategyCheck the original
Event sources
confirmed by 2 independent sources · 2 publishers, 2 independent. We count feeds from the same owner only once.