Apple plans to offer an AI server built around the “M8” series chips and has considered incorporating Nvidia networking hardware, The information reports.
The system would be sold to outside customers, potentially returning Apple to a business it left behind when it abandoned Xserve in 2011. Apple would target companies that want to run AI models on their own equipment, with a particular focus on inferring and generating responses from trained models.
One option under consideration is Nvidia’s NVLink Fusion, which provides hardware and software for exchanging data between chips. Apple could use the platform to link its M8 processors, although neither the Nvidia arrangement nor the server itself is finalized. The report sets a potential release date of 2029, but it could be canceled before then.
Apple CEO John Ternus reportedly supported the initiative when work began about a year ago, when he was in charge of hardware engineering at the company. The project follows growing interest in Apple hardware among AI companies, which are purchasing Mac mini and Mac Studio systems in large quantities.
Apple already produces server hardware for Private Cloud Compute, the infrastructure that handles Apple Intelligence tasks requiring processing beyond a user’s device. According to the report, scaling existing connections between Apple’s chips would pose cost and speed issues, which is why Apple is considering Nvidia’s technology. Apple has also declined requests from partners wanting to use these Private Cloud Compute servers.
A deal would give Nvidia a role in an Apple product that could compete with its own AI systems. NVLink Fusion allows Nvidia to provide networking technology to companies building alternative processors, expanding its business beyond customers using Nvidia chips.
The discussions also suggest a closer working relationship between Apple and Nvidia following their long-standing disputes. Apple has already announced plans to extend Private Cloud Compute to Google Cloud using Nvidia GPUs. The companies are reportedly exploring other ways to work together on AI, including using Nvidia’s open source models.
The project still faces challenges beyond hardware development. There is still a need for greater support for enterprise customers and more software resources for AI developers, including additional investments in its MLX framework.
