Supermicro has started shipping its new four-socket X14 servers, which it claims will enable enterprises to power intense AI and virtualization workloads.

Debuted back in April 2024, the data center servers feature Intel’s flagship Xeon 6 processors with Performance-core (P-core) SKUs, which the vendor said will help boost HPC, storage, and edge applications.

Charles Liang, president and CEO of Supermicro, said the X14s "provide up to 16TB of memory, and up to six double-width GPUs” to support intensive workloads.

The four-socket servers have been certified for SAP HANA and Oracle Linux, with workloads able to be scaled up in a single node without the latency of horizontally scaling across networks.

The 2U 4-socket iteration can host two double-width GPUs, while the X14 4U system can host up to six in order to support AI training and inferencing projects directly tied to enterprise database applications.

With up to 344 cores per system and support for both traditional enterprise databases and modern AI workloads, the servers aim to eliminate the need for separate specialized systems for different computing tasks.

The launch comes as enterprises increasingly seek to integrate AI capabilities with existing database and Enterprise Resource Planning (ERP) systems. Supermicro’s new offering looks to help organizations consolidate massive workloads on a single system rather than distributing them across multiple servers.

“With our Data Center Building Block Solutions (DCBBS) approach and worldwide production capabilities, Supermicro can now deliver complete data center solutions to customers in all geographies and is first-to-market for this segment,” Liang added.

“We are able to provide organizations with a wide variety of systems for all of their IT needs.”