Supermicro Launches New Hardware for AI Radio Networks

The platform merges radio and AI workloads into modular servers to improve efficiency across edge and core network sites.

Updated on Oct. 21, 2026 in Data Centers

Isometric editorial illustration of a modular enterprise server rack in a minimalist data center environment, representing industrial hardware infrastructure. — Supermicro has released new modular hardware platforms integrating NVIDIA and Nokia technologies to consolidate AI and radio network workloads for telecommunications carriers. AI Illustration. Upload story photo >

Live Poll

Will integrating AI into mobile network infrastructure lead to more reliable service for your household?

Supermicro has released new hardware integrated with NVIDIA Aerial RAN Computer platforms and Nokia radio technology for edge and core network deployments. The available systems support AI workloads and radio access network operations across various infrastructure form factors.

Why it matters

This hardware integration targets improved spectral efficiency and network automation, allowing carriers to consolidate radio and AI workloads onto shared, distributed infrastructure.

The ARC-Pro node delivers 1.6 PFLOPS of FP4 AI performance within a 429 mm chassis, while the 4U server configurations support up to 6 TB of DDR5 memory. These specifications allow for dense compute consolidation compared to traditional separate radio and server hardware.

The players

Supermicro

A manufacturer of high-performance server hardware known for modular, energy-efficient designs and broad integration with NVIDIA accelerators.

NVIDIA

A designer of graphics processing units and AI computing platforms that focus on high-throughput data processing for edge and cloud environments.

Nokia

A telecommunications infrastructure provider specializing in radio access network (RAN) technology and enterprise network solutions.

The details

The systems utilize NVIDIA Aerial RAN Computer platforms to run radio access network (RAN) functions—the hardware connecting mobile devices to the core network—alongside AI tasks on the same architecture. Supermicro offers varied form factors, including short-depth 2U chassis for distributed cell sites and high-density 4U servers designed for core network hubs. The ARC-Pro configuration specifically leverages NVIDIA Grace CPUs for processing and Blackwell Edition GPUs to handle heavy computational loads like autonomous network operations.

October 20-22, 2026: The ARC MGX 2U platform was displayed at NVIDIA GTC Berlin.
Late 2026: Planned customer testing for the 2U ARC-Pro system.
First half of 2027: Planned general availability for the 2U ARC-Pro system.

The Tech Race

This release follows the industry shift toward software-defined radio networks, aligning with the O-RAN initiative's goal of virtualizing network functions. By merging radio and AI workloads, Supermicro competes with other hardware vendors aiming to dominate the infrastructure powering automated cellular networks.

These solutions are available now for immediate integration by network operators and MSO locations. Telecom providers will likely deploy these units to enable new edge-based AI services, such as enhanced voice and vision applications, directly at the cell site.

The takeaway

The industry is moving toward dense, combined compute nodes that handle both traditional networking and intensive AI inference at the edge. Monitor the general availability release in the first half of 2027 to see if the planned 2U ARC-Pro units meet performance benchmarks in live field tests.