Supermicro Solutions for NVIDIA Aerial RAN Computer

Distributed networking sites handle radio functions and time-sensitive inference at the far edge, in locations where space, power, and latency constraints are tight. In addition to handling RAN workloads such as Service Management and Orchestration (SMO) and traffic routing, these sites are essential for low-latency AI services such as voice, vision, and Physical AI.

Supermicro’s short-depth, front I/O systems are designed to achieve maximum power-efficiency and reducing network TCO, making them ideal platforms to run NVIDIA ARC at cell sites.

Aggregation sites and Mobile Switching Offices consolidate multi-cell traffic for coordinated scheduling, beamforming, pooled RAN-plus-AI processing, and higher-throughput inference. With greater GPU density and memory, these locations can run larger models and support model updates and retraining closer to the network edge.

Supermicro delivers ARC MGX 2U systems such as SYS-222GL-NR and ARC MGX 4U high-density platforms, giving multi-service operators scalable capacity from aggregation to core-hub AI workloads.