The MGX™ 4U AI Server is a high-performance AI and HPC-optimized platform designed to support large-scale training and inference. Featuring up to 8× dual-slot GPUs and powered by dual AMD EPYC™ 9005 Series processors, it delivers extreme compute density, high-bandwidth interconnects, and flexible expansion for next-generation workloads.
Extreme GPU Performance
- Supports up to 8× dual-slot GPUs (NVIDIA L40S, RTX Pro 6000, or H200 NVL)
- NVIDIA® NVLink® Bridge support for high-bandwidth GPU interconnect
High Compute Power
- 2-socket AMD EPYC™ 9005 Series processors, up to 5GHz
High Scalability and Expansion
- Up to 5× PCIe 5.0 x16 slots for NICs
- Up to 8× E1.S SSD for high-speed storage
