Ever since the first deployment of 5G networks, carriers have been searching for a way to monetize the billions that they have spent on infrastructure and spectrum. The Killer App, if you will. Providing AI services may be the answer to their needs for additional revenue. Telecoms like SK Telecom (NYSE: SKM) have made no secret about their ambition to become an AI company.
During MWC25 in Barcelona, Spain earlier this month, SK Telecom unveiled an AI data center initiative, the “AI Pyramid 2.0 Strategy,” which includes an AI data center business model.
“We will advance customized AI data center products that can meet all types of customer demand, and AI agents B2B and B2C will also achieve AI sales growth by innovating services that provide new experiences to customers,” Yoo Young-sang, president of SK Telecom, said.
In 2023, SK Telecom launched its “AI Pyramid Strategy,” which promised to use AI to transform its core business areas, including mobile, broadband and enterprise, as well as new business areas like mobility and healthcare. This effort centers on providing AI Infrastructure, AI Transformation and AI Service. Using AI Infrastructure, which consists of data center, AI semiconductor, and multimodal large language models, SK Telecom has fashioned itself as an AI company.
Earlier this year, SK Telecom launched the on-demand SKT GPU-as-a-Service, which allows businesses to choose the number of GPUs and duration use as part of a customized AI service.
SK Group, which owns SK Telecom, is going one step further designing eight modular data centers to house AI infrastructure, including GPUs and neural processing units, in a space the size of a freight container. The data centers are expected to generate demand among startups and research institutions that need to quickly acquire cost-effective, low-capacity data centers.
In January, Verizon Business announced Verizon AI Connect, which is designed to deliver AI workloads for macro 5G networks, fiber connectivity and edge compute environments. Google Cloud and Meta are already buying additional capacity from Verizon to support their AI workloads.
“We are seeing significant demand for reliable network infrastructure that can support existing AI workloads,” said Kyle Malady, CEO, Verizon Business.
A shift toward real-time decision-making — known as inferencing — is poised to drive massive demand for additional computing power to the edge. More than 60 percent of AI workloads is expected to shift to real-time inference by 2030, according to McKinsey & Company. This will boost the need for low-latency connectivity, compute and security at the edge, .
The convergence of AI and edge compute facilitates the execution of AI tasks, including machine learning, computer vision and natural language processing, according to Nbsys, a provider of AI-driven 5G and LTE communications systems specializing in integrating advanced technologies across multiple sectors.
“The technologies of 5G and edge computing exhibit distinct mutually beneficial characteristics and serve different purposes,” Nbsys writes. “The implementation of 5G technology results in enhanced data transfer speed and capacity, whereas the utilization of edge computing mitigates the latency and bandwidth demands associated with data processing.”
AI coupled with 5G will profoundly enhance industry efforts in the future in the areas of autonomous vehicles, industrial automation, remote surgery and empowering IoT ecosystems.
“The integration of 5G’s rapid data transfer capabilities with edge computing’s efficient data processing capabilities can facilitate novel opportunities for real-time applications, improved digital encounters and empowered IoT ecosystems,” Nbsys writes.
