Two of the big themes at the 2024 edition of Mobile World Congress in Barcelona this week are artificial intelligence and 5G — and although one might view the two as separate technology trends, they are linked.

Service providers are currently investing heavily in maintaining and upgrading their services to comply with evolving tech standards for mobile networks. Despite efforts to improve service quality and capabilities, telcos have struggled to increase revenue, creating ongoing financial challenges and causing many to focus on cost savings. At the same time, AI requires a heavy investment in infrastructure, which service providers may be hesitant to do until they can be sure there’s a strong return on investment.

This divergence presents an opportunity for the telecom industry to integrate AI into its operations and has set the stage for creating the AI-RAN Alliance. The new initiative, launched at MWC24, aims to bridge the gap between traditional telecom infrastructure and AI. It’s a consortium of leading companies, universities, and industry players, including Amazon Web Services Inc., Arm Holdings Ltd., DeepSig Inc., Ericsson, Microsoft Corp., Nokia Corp., Northeastern University, Nvidia Corp., Samsung Electronics Co. Ltd., SoftBank Group Corp. and T-Mobile USA.

The AI-RAN Alliance represents a strategic shift toward embracing AI and its growth potential to mitigate the telecom industry’s financial pressures, Ronnie Vasishta, senior vice president of telecom at Nvidia, explained during a launch briefing. The primary goal is to merge AI with radio access network technologies and create new business opportunities in telecom, particularly with the upcoming 5G and 6G networks.

Specifically, the alliance will focus on three key areas:

• AI for RAN focuses on the shared infrastructure that supports AI and the RAN. AI can be used to make the radio network more efficient.
• AI on RAN is to enable service providers and their ecosystem to create new applications that can run on 5G networks.
• AI for RAN, where AI is used to improve the spectral efficiency of running the network.

This alliance will jointly develop use cases, white papers, blueprints and guidelines for best practices and liaise the findings with existing standards organizations.

“One of the biggest challenges has been the availability of data. AI requires data to be effective, but data availability has been somewhat restricted,” Vasishta said. “So we want to create an environment where data can be shared, and datasets can be provided to enable AI to be more efficient. The verification testing will be conducted within the AI-RAN Alliance Labs, which members can access.”

Another challenge is enhancing 5G infrastructure by addressing the connectivity issues that have limited its potential. To achieve this, the infrastructure needs to be moved closer to the point of use, and software-defined RAN needs to be enabled on the same infrastructure. This will improve the user experience and service efficiency for current applications and allow advanced AI-driven apps like AI inference to perform better. By supporting low-latency, high-bandwidth apps, telecom companies can maximize the potential of 5G networks.

Mohamed Awad, senior vice president and general manager of the infrastructure line of business at Arm, echoed the sentiment, stating that AI could improve network infrastructure performance at both the level of individual operators and the overall system by bringing network equipment closer to the users. Awad also noted that using AI within RAN may improve other areas.

“This is one of those things where you’ve got a big problem, and the more teams you have looking at it and helping solve it, the better,” said Awad. “As we think about new capabilities that are being unlocked, this is really about what happens as the infrastructure matures and additional technology is added.”

Service provider success is critical to the health of the tech sector. Over the past several decades, the telcos have been continually marginalized to the point where most of their services are considered commodities. With the rise of cloud, mobility, and AI, the world has become network-centric, putting service providers in their strongest position for a long time.

One of the more interesting opportunities for telcos is AI inferencing at the edge, which can be used for healthcare, robotics, warehouse automation and more. This requires high bandwidth, low latency and highly distributed computing, all of which service providers are good at.

However, if telcos continue to get squeezed on prices and see margins decline, that creates a vicious cycle where they pull back on investing, which holds back innovation. This alliance focuses on helping service providers use AI to not only cut costs but also to identify those mythical new revenue streams that have eluded them for so long.

Zeus Kerravala is a principal analyst at ZK Research, a division of Kerravala Consulting. He wrote this article for SiliconANGLE.

Image: geralt/Pixabay