T-Mobile and Ericsson Complete AI Beamforming Test

en.Wedoany.com Reported - On October 20, U.S. telecom operator T-Mobile announced the results of an artificial intelligence beamforming live network test conducted with Sweden's Ericsson in New York. The two parties deployed AI-managed beamforming software on some commercial 5G Advanced base stations, achieving a 16% increase in downlink user throughput, a 12% improvement in spectral efficiency, and a 10% growth in downlink data transmission volume during the test period. The data came from actual mobile network traffic tests, with the highest performance gains occurring during peak business demand hours.

The test used Ericsson's radio access network compute platform and dedicated chips, analyzing T-Mobile's live network data through supervised learning models to dynamically adjust sounding reference signal resource allocation. The software optimizes beamforming and radio resource scheduling based on changes in the radio environment and network load. The test showed that the use of advanced beamforming with full-band sounding reference signals increased by approximately 50%.

T-Mobile activated the feature on a selected group of base stations in New York, implementing the test through station-by-station software upgrades. The company did not disclose the number of base stations participating in the test and plans to extend the trial to other markets later in the fourth quarter of 2026, with the goal of larger-scale deployment in the first quarter of 2027.

The two parties previously conducted an AI-native link adaptation software test on a commercial 5G Advanced network in May 2026, recording results of nearly 10% improvement in spectral efficiency and up to 15% increase in downlink throughput. The New York test used the AI-managed beamforming feature and represents another radio access network optimization technology validation by the two parties.