Qualcomm Technologies introduced a portfolio of AI-driven radio access network (RAN) innovations aimed at accelerating network autonomy and preparing operators for the transition to AI-native 6G architectures. The announcement includes the launch of an Agentic RAN Management Service within Qualcomm’s Dragonwing RAN Automation Suite, alongside new AI capabilities for commercial radio unit (RU) and distributed unit (DU) platforms. The company says the technologies deliver immediate performance and operational benefits for mobile network operators while establishing a framework for autonomous networking.
The Agentic RAN Management Service introduces an “agentic layer” built on collaborative RAN AI agents designed to monitor network conditions, analyze operational issues, generate response plans, and execute changes through closed-loop automation. Each AI agent specializes in specific operational domains and leverages a toolkit that includes rApp marketplace integrations, quality-of-experience analytics, digital twin simulation environments, and advanced network analytics. Qualcomm said the architecture operates across multi-vendor, multi-generation, and multi-topology networks and already builds on deployments of the Dragonwing RAN Automation Suite with Tier-1 operators worldwide.
Qualcomm also introduced a set of production-ready RAN AI capabilities designed to operate on existing commercial hardware platforms, allowing operators to deploy AI-driven optimization without requiring new radio infrastructure. The new features support improved coverage, spectral efficiency, operational efficiency, and faster network deployment cycles for current 5G networks while aligning with future 6G system architectures.
• Agentic RAN Management Service: Introduces collaborative AI agents that monitor networks, analyze conditions, generate action plans, and autonomously implement changes through closed-loop network management.
• Autonomous Demand-Based RAN Adaptation: Dynamically optimizes network resources based on traffic demand, including support for fixed wireless access deployments.
• Autonomous RAN Supervision: Uses AI to detect operational anomalies and coordinate corrective actions across heterogeneous network infrastructure.
• RAN Energy Management: Applies AI-driven optimization to reduce power consumption while maintaining performance targets.
• AI-Driven Uplink Adaptation: Runs on the Dragonwing QDU100 Telco Server Platform and uses machine learning to dynamically adjust uplink parameters, improving coverage, spectrum efficiency, and reliability under varying traffic conditions.
• AI-Driven Downlink Beamforming Channel Prediction: Available on Dragonwing QRU100 Radio Units and QDU100 platforms, enabling predictive beamforming that improves throughput, mobility performance, and massive MIMO efficiency.
• AI-Driven Factory Calibration: Uses machine learning to automate and optimize radio calibration during manufacturing, accelerating deployment timelines and improving hardware consistency.
“ Our Agentic RAN Management Service and expanding RAN AI portfolio give operators the ability to unlock measurable performance, efficiency, and cost improvements in the near term, while laying the critical foundation for enabling fully autonomous, AI-native networks on the path to 6G,” said Durga Malladi, Executive Vice President and General Manager, Technology Planning, Edge Solutions & Data Center, Qualcomm Technologies.
🌐 Analysis: Qualcomm’s announcement reflects the broader industry push to embed AI directly into network control loops as operators prepare for AI-native 6G architectures. Vendors including Ericsson, Nokia, and Samsung have also introduced AI-driven RAN automation capabilities over the past year, often integrating digital twins, rApps, and agent-based orchestration frameworks to support closed-loop optimization. Qualcomm’s approach emphasizes AI features that run on existing commercial RU and DU platforms, which may help accelerate near-term adoption without requiring large-scale hardware upgrades.