Apple is exploring a potential re-entry into the server market, a space it previously occupied with its Xserve rack technology until 2011. The company is currently developing new AI-focused M-series chips that may integrate NVIDIA networking technologies.
The proposed design involves utilizing the NVIDIA NVLink Fusion portfolio to interconnect multiple M8 Ultra chips. Specifically, the plan includes using NVLink for socket-to-socket communication and NVLink-C2C for die-to-die connectivity, while Apple focuses on the system-on-chip architecture.
The project is currently scheduled for a 2029 launch. This initiative follows the growing adoption of Mac mini and Mac Studio hardware among AI developers, with Apple reporting record sales of $10.4 billion for its Mac lineup in the last quarter.
The new server hardware is intended for enterprise and government clients. These organizations are already known to build custom clusters using existing Mac Studio units for local AI inferencing tasks, creating a foundation for Apple's potential return to the sector.
