KDDI and Ericsson Claim Gains with AI Uplink Field Trial

Why It Matters

The trial proves AI can materially boost uplink performance at scale, accelerating carriers’ path to fully autonomous networks and enhancing user experience for data‑intensive services. It also showcases an open rApp ecosystem that could spur localized innovation and faster global rollouts.

Key Takeaways

• AI-driven UIO rApp boosted 4G throughput by 9.6% in live trial
• 5G uplink SINR improved 27% using Ericsson’s EIAP platform
• Trial covered 1,500 5G and 1,300 4G cells across Japan
• UIO achieved Level 3.86 autonomy, moving toward Level 4 network automation
• Open rApp ecosystem lets Japanese firms develop and scale AI solutions globally

Pulse Analysis

The telecom industry is increasingly turning to artificial intelligence to manage the growing complexity of multi‑band, multi‑technology networks. Ericsson’s Intelligent Automation platform enables developers to package AI logic as reusable rApps, allowing operators to plug advanced analytics directly into the radio access layer. By focusing on uplink interference—a long‑standing bottleneck for both 4G and 5G—AI can dynamically re‑allocate resources, reducing latency and improving spectral efficiency without manual intervention.

In the KDDI trial, the Uplink Interference Optimizer delivered measurable gains: 4G throughput rose nearly 10%, 5G throughput increased 3%, and the 5G signal‑to‑interference‑plus‑noise ratio jumped 27%. These improvements were achieved while maintaining network stability, a critical factor for carrier-grade deployments. The system’s Level 3.86 autonomy rating indicates it can make near‑real‑time decisions with minimal human oversight, edging closer to the coveted Level 4 autonomous network where end‑to‑end optimization is fully automated. For subscribers, this translates to faster upload speeds, more reliable video calls, and smoother performance for emerging physical‑AI applications such as edge‑based analytics and autonomous vehicles.

Beyond the technical results, the trial underscores a strategic shift toward open, developer‑centric ecosystems. Ericsson’s EIAP developer portal and the broader rApp marketplace empower local firms to tailor AI solutions to regional market needs while leveraging a common platform for global scalability. As carriers worldwide grapple with spectrum scarcity and ever‑higher data demands, the ability to rapidly deploy AI‑driven optimizations could become a competitive differentiator. The success in Japan positions Ericsson to market the UIO rApp to other operators ahead of MWC 2026, potentially accelerating the industry’s march toward fully autonomous, AI‑optimized mobile networks.