SoftBank penned a collaboration with Samsung to expanded efforts in developing AI-augmented radio access networks (AI-RAN).

The agreement will see the pair research AI-RAN technologies, including systems for 6G. Samsung's base station implementation technologies will be paired with SoftBank’s ongoing AI-RAN research to develop systems focusing on the 7 GHz band.

The telecom giants plan to conduct field tests and explore potential use cases. This includes research into AI’s effectiveness at the physical layer (Layer 1) to see if the technology can help to better optimize deployments.

The pair also agreed to tap SoftBank's generative AI-powered Large Telecom Model to support their efforts. Unveiled back in March, the LTM was developed on SoftBank’s trove of network data in a bid to use AI to help design, manage, and operate cellular networks.

The telecom giants will also look to see if AI can help to improve resource utilization by integrating AI and RAN workloads on shared infrastructure. In what they describe as a “AI-RAN orchestrator,” Samsung and SoftBank aim to see if both technologies can operate efficiently on the same hardware.

“Through this collaboration with SoftBank we aim to define meaningful use cases for both operators and end users while securing key technologies for future commercialization,” JinGuk Jeong, EVP and head of Samsung’s Advanced Communications Research Center, explained.

SoftBank’s Samsung partnership is the latest in the carrier’s growing efforts to develop AI to improve its RAN offerings. It’s already working on separate deals with both Ericsson and Nokia to jointly build prototypes aimed at boosting network efficiency and performance.

Prior to signing up with Samsung, SoftBank unveiled a major stride toward using AI to boost RAN performance, employing a Transformer model – the same class of neural networks used to power apps like ChatGPT – to help deliver better throughput performance.

Simulations of its new transformer-based architecture that took place earlier this summer saw downlink throughput for terminals moving at 80 km/h improved by up to approximately 29%, a figure that rose to around 31% for a terminal moving at 40 km/h.

Meanwhile, long-time technology partner Nvidia is working with SoftBank on AI-RAN, leveraging the former’s AI Aerial accelerated computing platform to apply AI to 5G networks. Earlier this year, the pair field-tested an AI-RAN solution dubbed AITRAS at Nvidia's headquarters in Santa Clara, California.