Gayan Koralage, Director of Strategy at EDOTCO Group, explains how telecom infrastructure is evolving from fragmented connectivity assets into deeply integrated, full-stack digital platforms that combine connectivity, compute, and resilience within a single market.

The “Baobab Model”: Depth over breadth

The African symbol of life, the baobab tree, is not designed for expansion across borders; rather, depth within a single ecosystem is sustained, where water, energy, and life are anchored into one resilient structure.

A similar architecture is now being shaped in telecom infrastructure. Over the next five years, value creation is expected to shift toward full-stack, single-country platforms, where infrastructure is embedded deeply into national digital layers. These platforms are being aligned with sovereign data strategies, artificial intelligence adoption, and long-term digital roadmaps. Connectivity and compute are no longer treated as separate layers; they are being integrated into a unified infrastructure fabric combining towers, metro and last-mile fibre, indoor small cells, edge compute, power systems, and resilience frameworks.

This unified stack is being positioned beyond traditional telecom operators. Enterprises, hyperscalers, and OTT players are increasingly being served as primary customers. Demand is driven by low-latency compute, sovereign data processing requirements, and enterprise-grade network performance. As a result, infrastructure is being redefined to deliver both coverage and processing capability in a single layer. The convergence of connectivity and compute is emerging as the defining principle of the next telecom cycle.

Structural shift 1: From connectivity to token factories

Telecom architecture is being redefined from a connectivity layer into a data factory, and progressively into a token factory. Pricing constructs are expected to evolve beyond pure data consumption models toward hybrid structures combining GB usage with tokenized compute and AI-driven services.

Mobile users are being augmented by agentic AI, where autonomous agents execute transactions, workflows, and decisions continuously. Traffic patterns are shifting accordingly – machine-originated traffic is set to outpace human activity, with networks becoming uplink-heavy, latency-sensitive, and persistently active, independent of traditional peak cycles.

Global mobile data traffic is projected to nearly triple by 2030, while AI workloads are expanding at a significantly faster rate. These workloads require sub-10 millisecond latency, deterministic performance, and localised processing. Centralised architectures alone are insufficient; distributed edge layers are being introduced. Telecom infrastructure is therefore being positioned not just for transport, but as execution nodes for AI inference within a distributed compute grid.

Structural shift 2: Capital discipline in a high-cost environment

The macroeconomic environment has introduced a new level of discipline into infrastructure investment.

Since 2022, interest rates have increased by approximately 300 – 500 basis points. Power costs – a key component of tower operating expenditure - have risen by 25–70%, while construction related costs remain elevated.

These pressures have contributed to a reset in infrastructure valuations, with towerco multiples compressing from approximately 20-25x EBITDA at peak to 12 -16x in more recent transactions.

Despite this, demand for connectivity continues to expand. During periods of economic constraint, connectivity has consistently behaved as a “premium utility” - a dynamic observed during the COVID period and likely to persist as AI adoption accelerates.

In this environment, value creation is expected to be driven less by scale alone, and increasingly by efficiency, utilisation and yield per asset.

Structural shift 3: Compute moving to the edge

The previous investment cycle was dominated by centralised datacentre expansion. Today, more than 15 gigawatts of global capacity is under development, with Southeast Asia accounting for over three gigawatts of pipeline capacity.

However, constraints in power availability, latency requirements, and data sovereignty considerations are gradually shifting the architecture towards distributed compute.

AI training is expected to remain centralised. Inference, however, is increasingly being distributed.

Transporting large volumes of data across long distances introduces both latency and cost inefficiencies. As a result, compute is being positioned closer to where data is generated.

Fibre-connected tower grids are emerging as a natural foundation for this shift. In this model, towers are no longer simply endpoints of networks, but nodes within a distributed compute architecture.

The towerco reset: Back to fundamentals

There is a clear return to operational fundamentals with emphasis on efficiency, cost optimisation, and structural discipline. Energy management is becoming increasingly important, with solar hybridisation, battery optimisation, and intelligent energy analytics being deployed to stabilise margins. Leaner operating models are also being adopted to reflect higher cost-of-capital conditions.

At the same time, a broader strategic shift is taking place. Tower companies are gradually moving beyond passive infrastructure ownership towards a more integrated role within the digital ecosystem.

Owning the digital stack: Towers, fibre, and small cells

The shift is most visible in the expansion towards integrated digital infrastructure ownership within individual markets.

Fibre networks, particularly metro and last-mile segments, are being incorporated to anchor tower portfolio and enable high-capacity backhaul. This reduces reliance on third-party providers while opening access to enterprise connectivity opportunities.

Indoor connectivity is also becoming a critical layer. With approximately 80% of mobile data consumption occurring indoors, in-building coverage is increasingly central to network performance. Neutral-host small cell systems are being deployed across commercial and industrial environments, enabling scalable, shared infrastructure models.

Together, these layers form the foundation of a more integrated, full-stack infrastructure platform.

Connectivity meets compute: The edge opportunity

Within this integrated model, a subset of tower sites, typically 5–20% of portfolios, is being upgraded to support edge compute capabilities. These sites are characterised by strong fibre connectivity, stable power, and strategic location.

Micro datacentres and GPU-based inference nodes are being deployed to support localised AI workloads.

The economic model is complementary. Traditional tower tenancy continues to provide stable, long-term cash flows. Edge compute introduces an additional layer of revenue, typically higher yielding and often structured over longer tenures.

This creates the potential to increase revenue per site meaningfully, while improving overall asset productivity.

Resilience as infrastructure

As reliance on digital system increases, resilience is becoming a core requirement of infrastructure design.

Climate variability, energy constraints, and the critical nature of connectivity require networks that can maintain continuity under stress.

Telecom platforms are increasingly being enhanced with satellite backup, intelligent routing, and real-time analytics. This enables a form of resilience as a service, positioning telecom infrastructure as part of broader national critical systems.

The 2016–2032 cycle: From scale to intelligence

The evolution of telecom infrastructure can be observed across three phases.

From 2016 to 2021, the focus was on scale, with rapid tower deployment and coverage expansion. The subsequent phase introduced greater capacity, driven by fibre integration and datacentre growth.

The current phase, extending to 2032, is defined by integration and intelligence, where connectivity and compute are increasingly unified.

Value creation in this phase is expected to be driven not only by scale, but by functional depth, platform integration and improved asset utilisation.

Becoming the Baobab

The baobab model is defined by depth, resilience, and integration. Telecom infrastructure is now being reshaped along similar lines.

Platforms that combine towers, fibre, indoor networks, edge compute, power systems, and resilience layers within a single country are expected to carry increasing strategic importance and premium valuations.

The next phase of the telecom cycle is therefore unlikely to be defined by expansion across markets, but by integration within them.

The model is shifting: Depth will be prioritised over breadth. Connectivity will be combined with compute. Infrastructure will evolve into platforms.

The baobab, in this context, is less a metaphor – and more a direction of travel.

---

About the Author:

Gayan Koralage is Director, Group Strategy at EDOTCO Group, one of the world’s leading independent telecommunications infrastructure companies. He has been instrumental in EDOTCO’s evolution since its inception in 2013, contributing to its growth into a top-six global tower company with operations across multiple Asian markets.

Gayan is recognised as a thought leader in the telecom infrastructure sector, with a focus on the economics of 5G, network disruption, and the convergence of connectivity, compute, and digital infrastructure platforms. In his current role, he leads the Group’s strategic agenda, including long-term portfolio direction, pricing, and commercial value creation initiatives.