Catenaa, Tuesday, July 07, 2026- Samsung Electronics and Japanese telecommunications operator KDDI have reported strong results from a commercial trial of artificial intelligence-powered network optimization, with 5G download speeds improving by as much as 52% in dense urban areas.
The joint trial, which began in late 2025, deployed Samsung’s AI-based RAN Speed Optimizer (RSO) across KDDI’s live standalone 5G network. The technology delivered an average 31% increase in downlink throughput across the entire trial area while achieving peak improvements of 52% in heavily populated parts of Tokyo.
The results add momentum to the growing adoption of AI in radio access networks, or AI-RAN, as mobile operators seek higher network efficiency, lower operating costs and improved customer experience.
The trial covered hundreds of 5G cells operating on 100 MHz of spectrum in the 3.7 GHz time division duplex band.
Testing extended across urban, suburban and rural locations surrounding Tokyo, allowing Samsung’s AI models to learn from a broad range of real-world network conditions.
Unlike conventional optimization methods that apply identical settings across groups of cell sites, Samsung’s RSO individually configures each cell using artificial intelligence.
The platform analyzes environmental and network performance data automatically before recommending customized operating parameters for every cell site.
This cell-level optimization enables networks to adapt dynamically to changing traffic patterns, user density and radio conditions without requiring extensive manual intervention.
KDDI said the approach solves a long-standing industry challenge by making individualized radio tuning commercially practical.
Improved network efficiency translates into faster download speeds, lower congestion and more consistent connectivity for consumers.
The AI platform continuously adjusts network parameters to maintain performance during periods of fluctuating demand, benefiting applications such as video streaming, cloud gaming, web browsing and voice communications.
Operators also stand to reduce operational costs by automating network optimization tasks that traditionally required engineering teams to make manual adjustments.
Artificial intelligence is becoming a major focus across the global telecommunications industry as operators prepare networks for expanding data traffic driven by cloud computing, autonomous systems and AI applications.
AI-RAN combines machine learning with radio access network management to improve spectrum efficiency, reduce energy consumption and enhance overall network reliability.
Technology vendors including Samsung are investing heavily in AI-powered network management tools as operators increasingly seek software-driven methods to maximize returns on existing infrastructure.
Following the successful trial, Samsung and KDDI plan to evaluate additional commercial applications for AI-based optimization across broader areas of the mobile network.
The companies expect AI to play a growing role in supporting future 5G evolution and the eventual transition toward sixth-generation mobile networks.
Radio access networks represent the portion of a mobile network connecting user devices to the core telecommunications infrastructure through cellular base stations. Traditional optimization methods typically configure groups of neighboring cell sites using common parameter settings, limiting flexibility in complex environments. AI-RAN introduces machine learning models capable of continuously analyzing traffic patterns, radio conditions and environmental factors to optimize each cell independently. As operators worldwide face rising mobile data consumption and growing deployment costs, AI-driven automation is increasingly viewed as a practical method for improving network capacity, efficiency and user experience without extensive hardware upgrades.
