In a groundbreaking move, SoftBank and NVIDIA are teaming up to change how telecom networks operate by introducing a powerful innovation called the Artificial Intelligence Radio Access Network, or AI-RAN. AI-RAN isn't just about connecting people—it's about using artificial intelligence to reshape network infrastructure, opening doors for faster connections, and even creating new revenue streams for businesses in telecom.
So, what does this all mean? Traditional telecom networks are usually designed to handle peak loads but end up using only about one-third of their potential capacity. With AI-RAN, SoftBank and NVIDIA are finding a smart way to tap into that underused network power to handle AI-based tasks, like real-time data processing and inference for AI applications. Think of it like adding a side hustle to your network; now, the network can handle both 5G traffic and AI tasks without missing a beat.
In Japan’s Kanagawa prefecture, SoftBank recently demoed this AI-RAN solution powered by NVIDIA technology, showing the world how AI and 5G can operate on the same network simultaneously. The results? Impressive high-speed 5G performance while also running AI inference tasks, all thanks to the extra capacity on the network. This partnership is making artificial intelligence in business feel more tangible than ever, especially for telecoms.
According to SoftBank and NVIDIA, the financial benefits of AI-RAN could be huge. They project that for every $1 invested in this new AI-powered technology, telecom companies could earn up to $5 from offering AI-based services. This could mean a revenue boost of up to 219% per AI-RAN server added, a big incentive for telecom providers to adopt this new tech.
Read More: How AI-RAN is Changing the Future of 5G
SoftBank's demo went beyond just showing AI-RAN's potential; they demonstrated some impressive real-world applications. Using NVIDIA AI Enterprise, SoftBank ran various AI-powered applications like remote support for autonomous vehicles, robotics control, and localised, automated data retrieval—all right at the network's edge. The AI applications operated seamlessly, showing that AI-RAN isn’t just theoretical; it’s ready for real-world AI applications in business.
One of the highlights was SoftBank’s software-defined 5G radio stack, specifically designed to run on NVIDIA’s AI computing platform. This setup, including the L1 software developed by SoftBank, uses NVIDIA’s Aerial CUDA-accelerated RAN libraries. Looking forward, SoftBank plans to add NVIDIA Aerial RAN Computer-1 systems to their setup, which could reduce power usage by around 40% compared to traditional network infrastructure—great news for businesses looking to cut costs while boosting performance.
A standout feature of AI-RAN is its dynamic response to changing demand. In simpler terms, AI-RAN adjusts its computing power to meet the network’s needs in real-time, keeping everything running smoothly. This flexibility is a game-changer for telecoms, offering new applications of artificial intelligence that adapt based on network load. And with NVIDIA AI Enterprise’s serverless APIs and SoftBank’s custom orchestrator, the setup can automatically allocate resources, ensuring no computational power goes to waste.
This system, known as “AITRAS” by SoftBank, represents years of research in artificial intelligence, marking a five-year collaboration with NVIDIA. SoftBank’s Research Institute of Advanced Technology VP Ryuji Wakikawa highlights AITRAS as a pioneering step, showing how AI and RAN workloads can be managed in a way that’s both efficient and forward-thinking.
Read More: Revolutionizing Telecom with AI
The collaboration between SoftBank and NVIDIA could redefine the telecom industry by bringing artificial intelligence into the heart of network infrastructure. This shift is not just about improving performance but about introducing a new AI technology model that merges AI and telecom. The potential for AI-RAN to transform telecommunications could be the first of many human-AI collaborations across industries.
With AI-RAN, telecom operators may have the chance to create new AI-driven business applications, giving them an edge in a competitive market. And with NVIDIA’s support, SoftBank is setting a standard that other telecom providers may soon follow.
By expanding artificial intelligence applications in business, from supporting autonomous vehicles to improving data management and robotics control, AI-RAN is paving the way for smarter, more efficient digital transformation in telecom. As the AI and telecom worlds continue to overlap, we’ll likely see more innovations like AI-RAN that bridge the gap between technology and communication.
By clicking "Accept", you agree to the storing of cookies on your device to enhance site navigation, analyze site usage and assist in improving your experience.
