Nvidia, SoftBank claim AI, 5G telecom pilot first
AI processor leader joins Japanese telco on AI Aerial platform for outdoor pilot of what is said to be world’s first combined AI and 5G network
By
Published: 15 Nov 2024 10:45
As part of its ambition to bring the power of artificial intelligence (AI) to virtually every business, industry and consumer in Japan, and in the process open AI revenue streams potentially worth billions of dollars, Nvidia is claiming to have successfully piloted with telco
SoftBank the world’s first combined AI and 5G telecom network using Nvidia’s AI Aerial accelerated computing platform.
At the heart of the project is SoftBank’s aim to build Japan’s most powerful AI supercomputer using the
Nvidia Blackwell platform, with plans to use the Nvidia Grace Blackwell platform for its next supercomputer.
Using Nvidia AI Enterprise software,
SoftBank is also aiming to create an AI marketplace that can meet the demand for local, secure AI computing. SoftBank sees the new service, which supports AI training and 
edge AI inference , as positioning itself to become the AI grid for Japan, facilitating new business opportunities for the creation, distribution and use of AI services across the country’s industries, consumers and enterprises.
In the field of
AI-based radio access networks (AI-RANs), the two companies believe they have achieved a technology milestone with a comms network that can run AI and 5G workloads at the same time. Traditional telco networks are designed to handle peak loads and, on average, said the two firms, have used only one-third of that capacity. With the common computing capability provided by AI-RAN, it’s expected that telcos now have the opportunity to monetise the remaining two-thirds capacity for AI inference services.
In the outdoor trial conducted in the Kanagawa prefecture, SoftBank demonstrated that its Nvidia-accelerated AI-RAN service has achieved carrier-grade 5G performance, and was able to do so while using the network’s excess capacity to run AI inference workloads concurrently.
For the trial, SoftBank used Nvidia AI Enterprise to build real-world AI inference applications, including autonomous vehicle remote support, robotics control and multimodal retrieval-automated generation at the edge. All inference workloads were able to run optimally on SoftBank’s AI-RAN network. SoftBank’s fully software-defined 5G radio stack is optimised for Nvidia’s AI computing platform, and includes L1 software enhanced by SoftBank based on Nvidia Aerial Cuda-accelerated RAN libraries.
Nvidia and SoftBank added that the infrastructure has broad ecosystem support from the comms industry, offering operators the ability to transform base stations from cost centres into AI revenue-producing assets. They estimate that operators can earn roughly $5 in AI inference revenue from every $1 of capex it invests in new AI-RAN infrastructure. Taking into account its own opex and capex costs, SoftBank estimates it can achieve a return of up to 219% for every AI-RAN server it adds to its infrastructure.
AI processor leader joins Japanese telco on AI Aerial platform for outdoor pilot of what is said to be world’s first combined AI and 5G network.
www.computerweekly.com
www.forbes.com
www.electronicdesign.com
www.zhi-hsin.com.tw
www.apollowave.co.jp
ailabsinc.com
