Expectations are high that AI model learning in next-generation communications such as 5G and 6G will dramatically improve existing performance, increase data processing, and quickly build smarter and more efficient networks in the future.

Samsung Electronics, NVIDIA, and Keysight Technologies have joined hands to develop AI-RAN technology for next-generation 5G-Advanced and 6G networks.

They made a splash at MWC 2025 by demonstrating how they can use AI to overcome challenges in existing radio access networks (RANs) and dramatically improve their performance.

Traditional RANs have been constrained in throughput and resource efficiency, and have limitations in terms of low signal-to-noise ratio (SNR), especially when user equipment (UE) is far from the base station.

Accordingly, AI technology is receiving new attention. Leveraging AI can enable more precise channel estimation, optimize resource allocation, and reduce power consumption.

With Keysight’s channel emulation solution, Samsung Electronics experimentally implemented an AI channel estimation model, and combined with NVIDIA’s Grace Hopper superchip platform, successfully improved data throughput at cell boundaries by more than 30%.

This has proven that it can provide higher network efficiency and user experience.

“AI-based technologies can greatly enhance the flexibility and efficiency of networks, and this collaboration is an important stepping stone toward AI-native networks,” said Soma Velayutam, vice president of NVIDIA.

Charlie Chang, Executive Director of Samsung Research, emphasized the importance of the project, saying, “Through cooperation with the AI-RAN Alliance, we will contribute to the development of sustainable next-generation networks.”

“This collaboration will significantly contribute to improving performance and energy efficiency,” said Gianpaolo Tardioli, Vice President of Keysight, adding that he is positive about the impact AI will have on innovation in the communications industry.