Operators currently face significant challenges with traditional RAN, including limited throughput and high latency, which particularly impact uplink performance at cell edges.

AI modeling plays a crucial role in this initiative, allowing for precise channel estimation, optimized resource allocation, and reduced power consumption, all of which enhance overall user experience and network capacity.

The collaboration aims to improve throughput and expand system capacity within Samsung's virtual RAN (vRAN) software solution, while also focusing on energy efficiency.

Keysight Technologies has teamed up with Samsung and NVIDIA to develop advanced AI models aimed at enhancing 5G-Advanced and 6G technologies, particularly focusing on improving Radio Access Network (RAN) performance through AI-based channel estimation.

Soma Velayutham from NVIDIA highlighted that AI enhances the flexibility and efficiency of radio signal processing, leading to significant performance gains in wireless networks.

This collaboration seeks to tackle critical issues in commercial 5G deployments by increasing system capacity, improving network efficiency, and reducing power consumption.

Giampaolo Tardioli from Keysight emphasized the transformative potential of AI-native networks, particularly in enhancing performance and energy efficiency.

The implementation of this project utilizes Samsung's radio point and distributed unit, powered by NVIDIA's AI Aerial platform and the NVIDIA GH200 Grace Hopper Superchip, alongside Keysight's Channel Emulation and Core Emulation Solutions.

As part of the AI-RAN Alliance, this collaboration has already demonstrated a notable 30% improvement in cell edge throughput compared to traditional static approaches.

Executives from Samsung, NVIDIA, and Keysight have underscored the significance of this initiative in advancing next-generation telecommunications and paving the way for AI-native networks.

Lab tests have demonstrated that using AI for channel estimation can lead to a 30% increase in cell edge throughput compared to conventional methods, showcasing the effectiveness of this collaboration.