Ericsson and KDDI boost 5G uplink by 27% with AI optimization

Ericsson and KDDI have announced the successful completion of a large-scale field trial in Japan on May 11, 2026, demonstrating the power of AI in optimizing 5G uplink performance. The trial deployed Ericsson’s AI-driven Uplink Interference Optimizer (UIO) rApp, a software application that runs on the RAN Intelligent Controller (RIC) platform. By analyzing real-time network data and adapting radio parameters autonomously, the UIO rApp achieved a 27% improvement in Signal-to-Interference-plus-Noise Ratio (SINR) across live 5G sites. This is a significant gain for uplink-heavy applications such as video streaming, remote monitoring, and industrial IoT, where interference often degrades user experience.

The trial also advances the industry toward Autonomous Networks Level 4, a stage where networks can self-optimize with minimal human intervention. Ericsson’s UIO rApp is part of its broader portfolio of rApps designed to close the loop between network analytics and automated actions. KDDI’s involvement validates the solution’s readiness for commercial deployment in dense urban environments. The results suggest that AI-driven interference management can reduce operational overhead while simultaneously improving quality of service. As 5G networks grow more complex, such closed-loop automation will be critical for operators to maintain performance and cost efficiency.