As global demand for AI computing capacity accelerates, telecom operators are increasingly positioning themselves as owners of the physical infrastructure supporting it.
SK Telecom announced SK Horizon in August: a new dedicated AI infrastructure business that combines AI data centre assets and submarine cable operations under one entity. The new company was backed by ₩3.08 trillion (US$2.2 billion) in investment from firm KKR and the IMM Investment-Stonebridge consortium and is designed to expand Korea’s AI data centre infrastructure.
The move reflects a broader strategic shift underway across the telecom sector. After focusing on how AI can improve network operations, customer care, and process automation, many are now looking at AI as its own infrastructure to develop.
According to the EY-Parthenon Telco of Tomorrow study released earlier this month, 97% of telecom C-suite leaders anticipate AI will provide “major productivity and operational gains within five years.”
The study showed that operators are increasingly looking beyond connectivity for future growth opportunities, with cybersecurity, sovereign cloud, AI infrastructure and GPU-as-a-service emerging as key priorities. However, 80% of respondents said digital infrastructure is still a core asset that should be owned and managed by telcos.
There are still obstacles holding back the convergence of telecoms and data centres. Of the respondents, 53% said budget constraints held them back, and 43% cited insufficient skills in the area. The study also indicated that the executives were still experiencing a lack of alignment between AI technology and their business strategies.
Regardless, the shift is becoming increasingly visible in investment strategies. In an analysis of the sector’s evolution, Telecom Review Asia pointed to telecoms operators such as Singtel, whose data centre Nxera business is expanding across Southeast Asia. Reliance Jio is similarly building AI data centre campuses across India.
These companies join SK Telecom as firms building out digital infrastructure portfolios while also minding their core network assets.
What’s causing the shift?
The union of telecoms, data centres, and subsea cables is being driven by the growing requirements of AI workloads.
As hyperscalers build larger AI campuses to meet the compute demands of their training and inference models, the importance of linking those sites through high-capacity fibre and undersea networks continues to climb. For telecom operators, this creates a new opportunity.
Connectivity is still crucial but might increasingly be viewed as a lower-growth and lower-margin business. Ownership of data centres, cable landing stations and international subsea networks opens the telecoms companies up to higher-value parts of the AI supply chain.
“Competition is increasingly coming from outside traditional telecom boundaries,” commented Adrian Baschnonga, EY global telecommunications lead analyst, when EY’s report was released.
“Satellite providers are expanding their ambitions, while hyperscalers are simultaneously becoming both competitors and essential partners. This makes strategic positioning more important than ever as telcos determine where they can create distinctive value within increasingly interconnected ecosystems.”
The structure of SK Horizon can be seen as a noteworthy development. Converging connectivity and AI data centres follows the sign on the wall – that infrastructure is about managing the entire stack, rather than just focusing on compute.
There are implications here for the telecoms sector. Growing the industry may be driven less by subscriber numbers and coverage maps and more by ownership of the AI infrastructure. As this evolution takes place, new global telecoms leaders may emerge.
RELATED STORIES
Capacity Europe 2026
The 25th anniversary edition of Capacity Europe 2025 will bring together 3,500+ decision-makers from the global connectivity and digital infrastructure community.
