Apple reportedly building server packed with M-series Ultra chips for AI

If you're building AI models, Apple servers with M8 Ultra chips could offer a new hardware option when they launch in 2029.

Apple is reportedly developing enterprise servers powered by its M8 Ultra chips, specifically for AI workloads, with plans to offer them for sale to AI developers, businesses, and governments. The servers are expected to come in two configurations: one with two M8 Ultra chips and a larger version with four chips. Apple is also considering integrating Nvidia's NVLink Fusion technology to enhance chip communication within the data center. This project has reportedly been underway for about a year, with CEO John Ternus showing support. However, the plans are not finalized, and the servers are not anticipated to launch before 2029.

Apple's potential entry into the AI server market signals a significant move to leverage its custom silicon beyond consumer devices and into the enterprise data center space. This could introduce a new competitive dynamic to a market currently dominated by x86 architecture and GPU-centric solutions, potentially offering a more integrated and power-efficient alternative for specific AI development and deployment tasks. The consideration of Nvidia's NVLink Fusion also indicates a pragmatic approach, acknowledging the need for robust interconnectivity in high-performance computing environments.

The development suggests Apple aims to capture a segment of the growing AI infrastructure market, catering to those seeking optimized hardware for their AI workloads. While the 2029 timeline indicates a long-term strategy, it positions Apple as a potential challenger to established cloud providers and hardware vendors, particularly for on-premises or specialized AI deployments. This move could impact AI developers and businesses by offering more hardware diversity and potentially driving innovation in chip design and data center architecture.