Telecom operators are putting more money into AI infrastructure

For decades, telecom operators spent most of their capital on one thing: moving data from one place to another. Towers, fibre, and spectrum formed the backbone of the business. Yet gradually, that focus is starting to shift. Across many markets, operators are putting more money into AI infrastructure, driven by rising demand for computing power and tighter rules around where data can be stored and processed.

Traditional networks were built to transport traffic efficiently, not to handle heavy, constant AI workloads close to where data is created. Shipping data back and forth to distant cloud regions can add cost, delay, and regulatory risk. For some applications, that approach not works.

Industry research points to how widespread this shift has become. According to Omdia’s report, Telcos’ Strategic Investments in AI Infrastructure, operators worldwide are scaling capital investment in AI infrastructure to meet rising AI compute demand and sovereignty requirements that often favour local providers and domestic solutions. The pressures are pushing telcos to rethink how and where they invest, and what role they want to play in national and regional digital systems.

In practice, this means more spending on data centres, cloud platforms, specialised chips, and software that can support AI training and inference. In some cases, operators are expanding existing facilities.

In others, they are building new ones or upgrading edge sites so data can be processed closer to users, factories, or public services.

Sovereignty rules push compute closer to home

National and regional initiatives are reinforcing this direction. Programmes like the European Union’s planned AI Gigafactories are designed to build large-scale compute capacity in Europe. These efforts create openings for telecom operators to lead or co-lead major infrastructure projects, not handing them over entirely to global cloud providers. For governments concerned about control, resilience, and data location, local operators can look like a safer bet.

Data sovereignty remains a central driver. Governments are tightening requirements in sectors like healthcare, finance, defence, and public administration. Enterprises operating in these areas increasingly want assurance that sensitive data stays in national borders.

Telecom operators, already subject to local regulation and oversight, are well placed to offer that assurance when connectivity and compute sit under the same roof.

Early revenue signals, uneven paths forward

The shift is also starting to show up in revenue, even if it is still early. Omdia’s analysis points to signs of AI infrastructure monetisation taking shape. In South Korea, data centre services accounted for about 4% of SK Telecom’s revenue in the third quarter of 2025, with the operator targeting KRW 1 trillion in related revenue by 2030.

In the Middle East, Ooredoo expects digital infrastructure to contribute 12% of group revenue by 2030, up from 3% in 2025. The figures are small compared to core connectivity today, but they signal a change in mix.

Operators are funding this shift through multi-year capital commitments. Omdia finds that telcos in Asia, Europe, Canada, and the Middle East are investing heavily in cloud platforms, data centres, GPU-as-a-service offerings, and AI-driven radio access networks. For many groups, this involves long-term planning not one-off projects.

There is no single blueprint for how this is done. Approaches vary in structure, risk, and ambition. Some operators have set up dedicated subsidiaries to house digital and AI infrastructure, like STC’s Centre3 or Iliad’s Scaleway.

Others are pursuing joint ventures, as seen with Singtel. For groups like SK Telecom, SoftBank, and Ooredoo, AI infrastructure forms part of a broader reset, not a side business.

In some cases, this reflects a deeper re-balancing of capital. As returns from traditional connectivity remain under pressure, a smaller share of CAPEX is flowing into legacy network expansion, with more directed toward compute and digital platforms. That shift brings risk. AI infrastructure is capital-intensive, and demand projections can change quickly. Running these platforms also requires skills that many operators are still building.

Rather than trying to cover every possible use case, many telcos are focusing on areas where AI and network assets naturally intersect. These include private networks for industrial sites, video analytics, traffic optimisation, energy management, and fraud detection. In these settings, the link between local data, low latency, and AI output is easier to explain to customers.

As Julia Schindler, Principal Analyst, Strategy, at Omdia, puts it: “AI infrastructure has become a bet for telcos. The rapid growth of AI traffic, in combination with national sovereignty initiatives, creates a unique opportunity that more telcos will want to capture in the future.”

Whether that bet pays off will depend on execution. If AI infrastructure ends up underused, it will add to financial strain. If it supports clear demand and steady workloads, it could help operators move beyond pure price competition. What is clear is that the old model, where networks simply carried data to distant clouds, is starting to show its limits.

As AI becomes part of everyday business systems, telecom operators are adjusting their investment plans to keep intelligence closer to the network, and closer to home.

(Photo by David Arrowsmith)

See also: Europe’s digital infrastructure: Regulatory headwinds stall ambitions

Want to discover how IoT is transforming telecoms and connectivity? Join the IoT Tech Expo in Amsterdam, California, and London. Explore how innovations in 5G, edge computing, and IoT are shaping the future of networks and services. The event is part of TechEx and co-located with other leading technology conferences. Click here for more information.

Telecoms News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.

Tags: ai, connectivity, data centre, network, networks, operator, telecoms