Apple is all set to return to the enterprise server market as the rapid expansion of artificial intelligence drives demand for powerful, energy-efficient computing infrastructure.
According to The Information, the tech giant is developing an AI-focused server that could use its upcoming M8 Ultra chips. The proposed machines would target AI developers, businesses and government organisations that want to run AI workloads on their own infrastructure rather than rely entirely on public cloud services.
Apple is also considering two server configurations. One could feature two M8 Ultra processors, while a more powerful version could combine four chips. The reported project marks a major shift for Apple, which left the dedicated server hardware business after discontinuing its Xserve line in 2011.
The machines would primarily target AI inference, which involves running trained models and processing user requests. This differs from AI training, where systems require enormous computing resources to build and refine models. Apple's move would extend its custom-silicon strategy beyond Macs and other consumer devices into enterprise computing.
Apple is also evaluating Nvidia's NVLink Fusion technology to connect multiple M8 Ultra chips. NVLink Fusion combines networking hardware, chiplets and software designed to enable high-speed communication between processors inside data centres.
Apple engineers are reportedly considering the technology because its own connectivity solution may not offer the performance or cost profile needed for larger deployments.
A potential Apple-Nvidia collaboration would also be notable given the companies' historically strained relationship. However, the company did not make any official confirmation.
Also Read: Huawei Targets AI Chip Market with New 960DT, 960PR Processors for 2027