SoftBank, in collaboration with NVIDIA, has created an artificial intelligence radio access network (AI-RAN) that can simultaneously handle AI and 5G workloads.
The telecoms industry has largely embraced this new infrastructure, recognising its potential to optimise network functions and drive new revenue streams. A demo carried out in Japan’s Kanagawa prefecture showcased SoftBank’s AI-RAN solution, enhanced with NVIDIA technology, delivering high 5G performance while utilising excess network capacity for AI inference tasks.
Traditional networks are constructed to manage peak loads, typically using only about one-third of their potential capacity. With AI-RAN’s shared computing capabilities, telecom operators can now exploit the remaining two-thirds of their network’s capacity for AI inference services.
Estimates from NVIDIA and SoftBank suggest that for every $1 invested in AI-RAN infrastructure, telcos could potentially earn $5 in AI inference revenue. SoftBank estimates a return of up to 219% per AI-RAN server added to existing infrastructure, considering both operational and capital expenses.
Ronnie Vasishta, SVP of Telecom at NVIDIA, said: “Shifting from single-purpose to multi-purpose AI-RAN networks can mean 5x the revenue for every dollar of capex invested. SoftBank’s live field trial marks a huge step toward AI-RAN commercialisation with the validation of technology feasibility, performance and economics.”
AI-RAN real-world applications
SoftBank’s real-world applications of AI-RAN were also showcased during these trials.
Using NVIDIA AI Enterprise, SoftBank implemented various AI inference applications – such as remote support for autonomous vehicles, robotics control, and localised, automated data retrieval and generation – at the edge of the network. These inference workloads seamlessly ran on SoftBank’s AI-RAN system.
Among the technical achievements is SoftBank’s entirely software-defined 5G radio stack, which is finely tuned for NVIDIA’s AI computing platform. This setup includes L1 software enhancements developed by SoftBank using NVIDIA Aerial CUDA-accelerated RAN libraries.
Looking ahead, SoftBank plans to integrate NVIDIA Aerial RAN Computer-1 systems into its solution—anticipating a 40% reduction in power usage compared to traditional 5G network infrastructures.
Supply and demand
A crucial feature of SoftBank’s AI-RAN solution is its ability to dynamically adjust computing power in response to fluctuations in demand and supply, all while maintaining carrier-grade performance.
SoftBank aims to establish a robust ecosystem that aligns AI technology demand with supply, facilitated by NVIDIA AI Enterprise’s serverless application programming interfaces and SoftBank’s bespoke orchestrator. This will ensure that SoftBank can offload external AI inference tasks to an AI-RAN server whenever free computational resources are available.
Ryuji Wakikawa, VP and Head of the Research Institute of Advanced Technology at SoftBank, commented: “SoftBank’s ‘AITRAS’ is the first AI-RAN solution developed through a five-year collaboration with NVIDIA. It integrates and coordinates AI and RAN workloads through the SoftBank-developed orchestrator, enhancing communication efficiency by running dense cells on a single NVIDIA-accelerated GPU server.
“We are confident this AI-driven innovation, AITRAS, will pave the way for new business models in telecommunications, serving as a crucial factor in the transformation of mobile operators.”
(Photo by Igor Omilaev)
See also: Elisa and Nokia conduct Europe’s first 100G PON trial
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with other leading events including Intelligent Automation Conference, BlockX, Digital Transformation Week, and Cyber Security & Cloud Expo.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.
