matteogirelli - Fotolia

Opinion

Why AI monetisation, not 6G, is the real prize for telcos

In the first of a three-part series exploring the role of telcos in the AI economy, we examine why operators must look beyond next generation 6G networks and focus on capturing the value of AI workloads

By
  • Edwin Lin
Published: 17 Jun 2026

The telecom industry is once again investing in next-generation mobile networks. However, after the mixed monetisation outcomes of 5G, the salient question is no longer about what operators will build, but how they will monetise those investments.

In my earlier Computer Weekly columns, I explored how telcos are being squeezed between hyperscalers and systems integrators, and why infrastructure has become central to their relevance in the artificial intelligence (AI) era. The next phase of this debate is now emerging, shaped by three important signals.

First, Omdia’s recent work on AI monetisation highlights how aggressively operators are investing in AI-ready infrastructure – from datacentres and graphics processing units (GPUs) to edge computing platforms. The intent is clear: to position networks as the foundation for AI workloads.

Yet, the revenue reality remains modest. Telcos are expected to generate only around $4bn in AI revenue in 2025, even as growth accelerates rapidly from this low base. At the same time, AI is already reshaping demand, with up to a third of network traffic expected to be AI-driven.

Second, industry leaders such as David Soldani have argued that 6G must be operator-owned by design. The rationale is well understood: in previous technology cycles, telcos built the infrastructure while value accrued to those controlling software, platforms and customer relationships. Without greater architectural control, 6G risks reinforcing this imbalance.

Third, China is moving quickly beyond strategy and into execution. Its latest 6G initiatives reflect a coordinated approach that integrates networks, AI and industrial use cases from the outset, treating 6G not just as a connectivity upgrade, but as part of a broader AI economy ecosystem.

Taken together, these developments point to a clear shift. Telcos can no longer remain mere connectivity providers; they must reposition themselves as the infrastructure layer of the AI economy.

From infrastructure to value

This repositioning builds on trends already visible across the Asia-Pacific region. Operators such as Singtel, China Telecom and SK Telecom are investing heavily in AI-ready infrastructure to remain relevant in a hyperscaler-dominated landscape. In this context, infrastructure becomes a strategic lever.

However, becoming essential does not necessarily mean becoming profitable.

Despite growing demand for AI workloads across the network, the share of value captured by telecom operators remains limited. Hyperscalers continue to dominate the higher layers of the stack, accounting for roughly 65% of global cloud infrastructure spending in a market increasingly driven by AI.

The implication is a familiar one: telcos may host the growth of AI, but they will not necessarily capture its value.

Where value really sits

In the AI era, value is increasingly concentrated not in the network itself, but in what runs on top of it.

AI workloads – from training to inference – drive demand for connectivity and compute. However, the economics of AI are defined by platforms, developer ecosystems and enterprise applications. This is where pricing power resides, where customer relationships are anchored and where margins are highest.

The rise of hyperscaler marketplaces reinforces this shift. These platforms are rapidly becoming the primary route to market for enterprise software and AI services, pulling value even further away from the infrastructure layer.

At the same time, AI is reshaping the network itself. Inference workloads are becoming more distributed, moving closer to users and requiring low-latency execution. While this plays directly to telecom strengths, proximity to demand does not automatically translate into ownership of value.

From connectivity to outcomes

One response from the industry has been a shift from selling connectivity to delivering outcomes. Instead of simply selling bandwidth, telcos are focusing on performance metrics, such as latency guarantees, reliability and service-level agreements (SLAs), aligned to specific use cases. This is a necessary evolution, aligning telecom services more closely with enterprise needs and the requirements of AI-driven applications.

But it is not sufficient. As long as value creation is anchored in platforms, services and ecosystems above the network, telcos risk remaining structurally disadvantaged, regardless of how advanced their infrastructure becomes.

As the industry looks towards 6G, the debate around ownership is necessary, but incomplete. Ownership of infrastructure matters. Control of architecture matters. Yet, neither guarantees value capture.

The more important question is not who owns 6G, but who owns and monetises the AI workloads it enables. In an AI-driven economy, networks are not the endpoint; they are the foundation. And foundations, while essential, do not always lead to optimal value capture.

What comes next

If AI is reshaping both demand and value in telecoms, the next question is how that demand will be measured and how it will be priced.

Will it still be measured in gigabytes of data? Or will new models linked to compute, inference and AI usage emerge?

That is the question I will explore in part two on why telco economics must change.

Edwin Lin is principal consultant at Omdia, part of Informa TechTarget.

Read more on Mobile networks