Lenovo ThinkSystem SR665 V3 AMD EPYC Server

The Lenovo ThinkSystem SR665 V3 (7D9AA01HNA) is a 2U dual-socket rack server built on the AMD EPYC 9004 series (Genoa) processor architecture, delivering exceptional core density, memory bandwidth, and PCIe 5.0 I/O throughput for demanding enterprise workloads. Supporting two AMD EPYC 9004 series processors with up to 96 cores each, the platform provides a total of up to 192 physical cores per server node, making it one of the most compute-dense 2U platforms available for virtualization consolidation and parallel computing tasks.

The SR665 V3 supports up to 24 DDR5 DIMM slots per processor socket (48 total), enabling very large memory configurations suitable for in-memory databases, large virtual machine footprints, and analytics workloads. The platform incorporates PCIe 5.0 connectivity across multiple expansion slots, supporting high-bandwidth accelerators, NVMe storage, and high-speed networking adapters without bottlenecks. Lenovo's XClarity management integration provides comprehensive out-of-band management, firmware lifecycle control, and telemetry collection compatible with enterprise data center automation frameworks.

Designed for mixed enterprise environments, the ThinkSystem SR665 V3 is well suited for cloud-native infrastructure, high-performance computing clusters, AI inference and training with GPU acceleration, and software-defined storage nodes. Its flexible storage backplane options support both NVMe and SAS/SATA drive configurations, and its tool-less chassis design facilitates efficient deployment and maintenance in high-density rack environments. The server carries Lenovo's enterprise-grade engineering standards and is validated for major hypervisors, Linux distributions, and HPC software stacks.

Ideal for

• Large-scale server virtualization consolidation leveraging high per-socket core counts to maximize VM density per rack unit
• High-performance computing (HPC) cluster nodes for scientific simulation, computational fluid dynamics, and genomics workloads
• AI and machine learning inference or training deployments using PCIe 5.0 GPU accelerator cards
• In-memory database hosting (SAP HANA, Redis, Oracle) benefiting from large DDR5 memory capacity and bandwidth
• Software-defined storage nodes requiring high core counts for storage controller overhead and NVMe-oF target workloads
• Enterprise cloud infrastructure build-outs running OpenStack, VMware vSphere, or Red Hat OpenShift on high-density 2U compute nodes

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.