Samsung AI RAN trial with KDDI boosts 5G speeds by 52%
Samsung and KDDI have deployed an AI-powered RAN Speed Optimizer on a commercial 5G network to automate performance tuning at the cell level. This technology adjusts network settings in real-time to improve throughput and reliability, specifically targeting high-data applications like streaming and gaming.
Key Takeaways
- Peak-hour downlink throughput increased by an average of 31% across the trial area, reaching 52% in dense urban zones.
- The deployment replaces conventional cluster-level tuning, which applies uniform settings to groups of cells, with granular site-specific optimization.
- The RAN Speed Optimizer (RSO) is a core component of the Samsung CognitiV Network Operations Suite (NOS).
- Testing leveraged 100 MHz of TDD spectrum in the 3.7 GHz band across hundreds of live cells in the greater Tokyo region.
Why It Matters
This trial marks a shift from reactive, manual network maintenance to proactive cellular automation. For streaming providers, these capacity gains directly translate into reduced buffering and higher bitrate stability in congested urban areas without requiring telcos to deploy new physical hardware. The ability to optimize downlink throughput by over 50% through software indicates that virtualization and AI will be the primary drivers for 5G ROI. As 5G Standalone networks expand, watch for telcos to prioritize AI-driven load balancing as a way to manage the surge in 4K mobile video traffic. A key signal will be whether these gains can be replicated in multi-vendor Open RAN environments beyond Samsung's own stack.
Additional Context
The Samsung and KDDI trial, which concluded in mid-2026, builds on a strategic shift toward software-defined infrastructure that began accelerating in 2024. Per Samsung’s January 2025 assessment, virtualized RAN (vRAN) has matured to a point where operators are prioritizing integration and security over initial cost savings. This transition is essential for the industry’s progression toward 'AI-native' 6G, which is expected to support 30Gbps data rates and integrated sensing technologies. Samsung has already achieved single-user downlink speeds of 3.05Gbps in testbeds using 7GHz spectrum, highlighting the growing need for autonomous management as radio environments become increasingly complex. Simultaneously, other major infrastructure players are aggressive in the AI-RAN space. Per RCR Wireless in July 2026, while the 52% gain is a significant milestone, industry analysts caution that these metrics currently focus on downlink throughput. Vital performance indicators such as latency, handover success rates, and uplink speeds—critical for interactive applications like cloud gaming and live mobile broadcasting—have not yet been as publicly detailed in these deployments. Mobile operators are now seeking evidence of robust 'rollback' mechanisms and multi-vendor interoperability to ensure AI automation does not create new points of failure. Regionally, Japan continues to be a primary testing ground for advanced RAN architectures. Per ComputerWeekly in June 2026, KDDI’s early move to fully virtualized network layers provided the necessary computational foundation to ingest the real-time telemetry required for Samsung’s optimizer. This follows similar logic seen in the U.S. market, where Fixed Wireless Access (FWA) subscribers reached 11.5 million by the end of 2024, per Samsung data. The convergence of vRAN and AI is now viewed as the primary mechanism for operators to squeeze more value from their mid-band spectrum assets amidst these rising data demands.
Read full article at trendhunter.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source