ASUS RS740-E12-RS24U 4U AMD EPYC Dual-GPU Storage Dense Server

The ASUS RS740-E12-RS24U is a 4U rack-mounted server engineered around the AMD EPYC 9004 series (Genoa) processor platform, delivering exceptional compute density combined with expansive storage capacity and dual-GPU support. Built on the ASUS ESC4000A-E12 class architecture, this system accommodates two AMD EPYC processors, supporting up to 6TB of DDR5 registered ECC memory across 24 DIMM slots, and is designed to address the demanding requirements of modern AI inference, deep learning training, and high-throughput data analytics workloads in enterprise and hyperscale datacenter environments.

The RS740-E12-RS24U integrates a high-density storage subsystem with 24 hot-swappable drive bays, enabling organizations to deploy large-scale object storage, distributed file systems, or tiered storage architectures without sacrificing GPU compute capability. The dual GPU configuration supports full-length, full-height, double-width cards — including NVIDIA A100, H100, and L40S class accelerators — via PCIe 5.0 slots, making it suitable for co-located GPU compute and local NVMe or SAS/SATA storage pipelines that eliminate I/O bottlenecks common in disaggregated architectures.

Targeted at enterprise IT, telecommunications, financial services, and research institutions, the RS740-E12-RS24U ships with ASUS ASMB11-iKVM out-of-band management, Redfish-compliant BMC, and support for ASUS Control Center (ACC) for fleet-level provisioning and monitoring. Its 4U chassis balances airflow, thermal management, and power delivery through redundant 3000W platinum-rated PSUs, meeting the operational requirements of dense rack deployments across UAE, GCC, EMEA, and APAC datacenter facilities.

Ideal for

• AI inference and deep learning model training leveraging dual high-end GPU accelerators with direct local NVMe dataset access
• Software-defined storage deployments requiring high drive bay count combined with substantial on-node compute for metadata processing
• High-performance data analytics pipelines where GPU acceleration and large local storage eliminate cross-network data movement
• Genomics, scientific computing, and HPC workloads requiring large memory footprints and parallel storage throughput
• Edge AI and regional datacenter nodes consolidating GPU compute and storage in a single space-efficient 4U chassis
• Telco and financial services real-time data processing requiring low-latency local storage paired with accelerated compute

Technical specifications

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