Sources: Apple is developing enterprise AI servers powered by two or four M8 Ultra chips, set for 2029, and is considering integrating Nvidia's NVLink Fusion

Apple's custom AI servers could offer a new, powerful option for enterprise AI deployments, potentially disrupting the market for specialized AI hardware.

Apple is reportedly developing in-house AI servers designed for enterprise use, with a target release date around 2029. These servers are expected to be powered by either two or four of Apple's M8 Ultra chips, a configuration that suggests a significant leap in processing power for its silicon. In addition to its custom chip strategy, Apple is also exploring the integration of Nvidia's NVLink Fusion technology. This potential collaboration indicates Apple's intent to leverage high-speed interconnects to enhance the performance of its server hardware, particularly for demanding AI workloads.

Apple's strategic move into enterprise AI servers, leveraging its custom M8 Ultra chips and potentially Nvidia's NVLink Fusion, signals a long-term commitment to dominating AI infrastructure. This development could challenge established players by offering a tightly integrated hardware and software ecosystem, akin to its approach with iPhones and Macs.

The integration of NVLink Fusion suggests Apple aims to achieve server-level performance that rivals or surpasses existing solutions, particularly in distributed AI training and inference. This positions Apple not just as a consumer device maker but as a significant contender in the lucrative enterprise AI hardware market, potentially influencing future server architecture and chip design trends.