NVIDIA and Partners Show That Software-Defined AI-RAN Is the Next Wireless Generation

Source: NVIDIA AI Blog

AI‑RAN: From Lab to Field

Key Message: A software‑defined approach is the only viable way to build future AI‑native wireless networks.

Recent Announcements (MWC 2026, Barcelona – March 2‑5)

• NVIDIA & Nokia unveiled new AI‑RAN collaborations with leading telecom operators across Europe, Asia, and North America.
• Operators achieving milestones:
  • T‑Mobile (U.S.)
  • SoftBank
  • Indosat Ooredoo Hutchison (IOH) – these partners have taken NVIDIA‑powered AI‑RAN outdoors and over‑the‑air.

Performance Highlights

• Benchmarking results from partners such as SynaXG demonstrate that AI‑RAN on NVIDIA platforms delivers:
  • High‑speed, carrier‑grade performance (extreme reliability)
  • Consistent operation across multiple 5G spectrum bands

MWC Demonstrations

• 20+ demos built on NVIDIA platforms will be showcased at MWC, under the AI‑RAN Alliance.
• These demos illustrate how AI is:
  • Boosting 5G performance and efficiency
  • Enabling new edge‑AI applications

Looking Ahead

All of this momentum points toward a common, software‑defined foundation that will pave the way for secure, open, and AI‑native 6G systems.

AI‑RAN Goes From Lab to Live

Top telecom operators and partners are using NVIDIA platforms to bring AI‑RAN to commercial deployment.

T‑Mobile (U.S.)

• Demonstrated concurrent AI and RAN processing on the NVIDIA AI‑RAN platform using Nokia’s CUDA‑accelerated RAN software.
• In an over‑the‑air field environment, Nokia’s AirScale massive MIMO radio (3.7 GHz) supported commercial devices running video streaming, generative AI, AI‑powered video captioning, and 5G simultaneously.

SoftBank – AITRAS

• Completed an industry‑first live field trial of a 16‑layer massive MIMO system, fully software‑defined 5G on NVIDIA’s AI‑RAN platform.
• Marks a key technical milestone toward AI‑RAN commercialization.

IOH (Indonesia)

• Implemented software‑defined 5G with Nokia’s vRAN software on NVIDIA AI‑RAN platforms, moving from proof‑of‑concept to pre‑commercial field validation.
• Showcased at MWC: Southeast Asia’s first AI‑powered 5G call, enabling secure, real‑time cross‑border connectivity and remote control of a robotic dog over a live 5G network.
• Demonstrates readiness to scale AI‑native network capabilities across Indonesia.

SynaXG

• Demonstrated fully software‑defined AI‑RAN using NVIDIA AI Aerial – a suite of accelerated computing platforms, software libraries, and tools for building, training, simulating, and deploying AI‑native wireless networks.
• Ran 4G and 5G in both sub‑6 GHz (FR1) and millimeter‑wave (FR2) bands, alongside agentic AI workloads, on a single NVIDIA GH200 server – the world’s first AI‑RAN implementation on FR2 bands.

Performance Highlights

• Activated 20 component carriers with both a Centralized Unit (CU) and Distributed Unit (DU) on one platform.
• Achieved 36 Gbps throughput (video demo) and sub‑10 ms latency.
• Highlights AI‑RAN‑based 5G performance and seamless orchestration between AI and RAN workloads.

Tripled Pace of AI‑RAN Innovation

This year’s MWC will showcase three times the AI‑RAN innovations of last year. Out of 33 AI‑RAN Alliance demos, 26 are built with NVIDIA AI Aerial and a software‑defined architecture.

Highlighted Demos

• DeepSig – Reinventing how devices “speak” to networks.

  • AI learns a smarter signal format at both ends of the link, jointly encoding and decoding signals with neural techniques at the device and base station.
  • Removes pilot overhead and adapts to site‑specific channels.
  • Early NVIDIA‑platform results: up to 2× higher throughput and improved spectral & energy efficiency.
  • 👉 DeepSig AI‑native Open RAN demo

• SUTD, NVIDIA & partners – Split‑inference for robots and autonomous vehicles.

  • Distributes “thinking” across device, edge, and cloud.
  • Real‑time task placement meets strict latency, privacy, and coverage SLAs, enabling physical AI and vision‑language models to scale at the network edge.

• zTouch Networks & partners – AI‑RAN orchestration blueprint.

  • Demonstrates safe GPU sharing between AI and RAN workloads using NVIDIA Multi‑Instance GPU technology.
  • Real‑time resource steering maximizes GPU utilization, improves energy management, and preserves RAN QoS.
  • A key step toward commercial multi‑tenant AI‑RAN solutions that turn GPU capacity into revenue.
  • 👉 NVIDIA Multi‑Instance GPU

• Northeastern University & SoftBank – AI switching for NVIDIA AI Aerial.

  • Microsecond‑level switching between AI‑based and classic channel‑estimation algorithms.
  • Dynamically selects the optimal processing path, boosting stability, throughput, and demonstrating AI‑classic coexistence.

“AI‑RAN is emerging as a unifying architecture for future radio networks,” said Alex Choi, chair of the AI‑RAN Alliance.
“By aligning operators, vendors, and researchers around software‑defined, GPU‑accelerated architectures, we accelerate innovation, validate concepts quickly, and lay the groundwork for AI‑native 6G—today.”

AI‑RAN for Autonomous Mobility

• Capgemini (Project ULTIMO, Horizon Europe) showcases AI‑RAN‑enabled autonomous mobility across European cities.
  • Autonomous shuttles equipped with NVIDIA Jetson Orin process sensor data locally.
  • Selected video and telemetry streams are sent over 5G to agentic AI applications on NVIDIA AI‑RAN servers.
  • These workloads handle scene understanding, incident/safety detection, and accessibility insights at scale, while mission‑critical 5G traffic receives priority GPU access.
  • 👉 Capgemini AI‑RAN in Action

All demos illustrate how software‑defined, GPU‑accelerated RAN is rapidly becoming the backbone of AI‑native 6G and next‑generation autonomous services.

A Growing Ecosystem

A growing ecosystem of partners is forming around NVIDIA‑powered AI‑RAN platforms, giving operators a variety of deployment options. NVIDIA Aerial RAN Computer (ARC) platforms combine the NVIDIA Grace CPU with a range of GPUs to deliver a high‑performance, energy‑efficient foundation for AI‑native RAN infrastructure.

• Quanta Cloud Technology (QCT) – Announces commercial‑off‑the‑shelf AI‑RAN products that support NVIDIA ARC platforms and Nokia software, providing operators with standardized building blocks.
• Supermicro – Extends support across the full NVIDIA AI‑RAN portfolio, including ARC‑Pro, RTX 6000‑based configurations, and ARC‑Compact systems with Nokia software.
• WNC – Introduces a new AI‑optimized indoor/outdoor open radio unit integrated with the NVIDIA AI Aerial Testbed and ARC platforms, supporting 5G‑Advanced and future 6G use cases.
• Eridan – Launches a 4T4R O‑RU (alongside its 2T2R O‑RU) integrated with NVIDIA AI Aerial, plus a DU running on the NVIDIA DGX Spark desktop supercomputer, merging spectrally efficient radios with GPU‑based baseband processing for a portable outdoor base station.
• LITEON – Completes integration of its sub‑6 GHz and millimeter‑wave radio units with NVIDIA AI Aerial and expands collaboration with ecosystem partners such as Supermicro and SynaXG to accelerate AI‑RAN commercialization.

Laying the Foundation for Open, Secure, AI‑Native 6G

NVIDIA’s latest State of AI in Telecom report shows that the industry is accelerating AI‑native RAN and 6G investments—signaling a major inflection point ahead of the traditional 6G deployment cycle. 77 % of respondents anticipate a much faster time‑to‑deployment for this new AI‑native wireless network architecture.

This progress in software‑defined AI‑RAN is setting the stage for secure, open, and AI‑native 6G systems.

• NVIDIA has open‑sourced the NVIDIA Aerial CUDA‑accelerated RAN libraries, fueling rapid AI‑RAN innovation.
• NVIDIA has also joined the OCUDU (Open CU‑DU) Ecosystem Foundation, hosted by the Linux Foundation, to contribute to open‑source RAN software development and accelerate research and commercialization for next‑generation wireless networks.

Learn more

• Meet NVIDIA and partners at Mobile World Congress.
• Explore key insights from the State of AI in Telecom survey.