T-Mobile and Ericsson Trial AI-Native RAN Scheduler on Live 5G Advanced Network

T-Mobile and Ericsson moved an AI-native Scheduler with Link Adaptation into large-scale commercial trials on live 5G Advanced network traffic, reporting up to 15% higher downlink throughput and nearly 10% better spectral efficiency versus legacy rule-based methods.

The Ericsson software runs a neural network directly on Ericsson RAN hardware to predict fast-changing radio conditions in real time. The feature aims to improve link adaptation decisions in the radio access network, especially in high-demand areas or locations with challenging RF conditions.

The companies said the large-scale live network results matched earlier testing across limited geographies, supporting broader deployment work as T-Mobile and Ericsson expand the trial footprint. T-Mobile said the effort follows its nationwide 5G Advanced deployment in 2025.

• Ericsson AI-native Scheduler with Link Adaptation entered large-scale commercial trials with T-Mobile.
• Trials used live 5G Advanced network traffic.
• Reported gains: nearly 10% spectral efficiency improvement and up to 15% higher downlink throughput.
• The software uses a neural network running on Ericsson hardware.
• The feature supports real-time radio condition prediction and faster RAN decision-making.
• Target customer benefits include more consistent streaming, gaming, and video call performance during peak usage.

“Following our milestone as the first U.S. operator to deploy 5G Advanced nationwide in 2025, we’re continuing to push the boundaries of RAN innovation. Our work with Ericsson on AI-native Scheduler with Link Adaptation demonstrates how real-time, AI-driven optimization can enhance spectral efficiency and throughput while delivering a more consistent experience for customers at scale,” said Grant Castle, Senior Vice President of RAN Engineering & Emerging Technologies, T-Mobile.

🌐 Analysis: The trial shows how 5G Advanced gives operators a practical framework for adding AI into live RAN operations, not just management-layer automation. Ericsson and T-Mobile are positioning AI-native RAN software as a way to extract more capacity from existing spectrum while preparing the network for more programmable, autonomous optimization.

Addendum: AI-Native Features Enter the 3GPP Standards Roadmap

Ericsson and T-Mobile’s AI-native scheduler trial reflects a broader industry transition already underway inside the 3GPP standards process. While current deployments use proprietary vendor AI models, 3GPP Releases 18 and 19 establish the framework for AI-assisted and eventually AI-native radio access networks.

Release 18, the first formal 5G Advanced release, introduced foundational AI/ML studies for the RAN, including beam management, channel state prediction, positioning, and network optimization. Release 19 expands that work into more operational capabilities tied to scheduling, link adaptation, mobility optimization, and AI-driven air interface enhancements.

The longer-term objective is to create more autonomous and programmable radio networks capable of continuously adapting to traffic conditions, spectrum availability, and application requirements in real time. Vendors including Ericsson, Nokia, Samsung, Huawei, and others are already using proprietary AI techniques inside standards-compliant 5G systems as the industry moves toward more AI-native 6G architectures.