On day 1 of TowerXchange Meetup Europe 2025, David McKean, Director of BT's Tower Division, offered a compelling examination of the convergence between telecom towers and datacentres. Drawing on his over 20-year long industry experience with Crown Castle, Telstra, and now BT, McKean shared his learnings and strategic foresight into how towercos might evolve to meet emerging demands, particularly in AI infrastructure and edge computing.
Can towercos transition beyond simple colocation models to become holistic infrastructure providers?
The future of telecom infrastructure lies in a hybrid model where towers, datacentres, fiber, and edge nodes interlock seamlessly. For towercos, this represents both a challenge and an unprecedented growth opportunity.
Many towercos’ remits have expanded to explore asset-based adjacencies, with datacentres being an emerging one. There are comparative advantages, challenges, and emerging opportunities for towercos as they navigate the rapidly evolving digital infrastructure landscape. There are already towercos doing it - TDF owns and operates a network of data centers and "mini" datacentres in France, CETIN offer datacentre leasing and data services for corporate networks and Cellnex operates a portfolio of datacentres, including edge datacentres, in several countries.
Towers vs. Datacentres
Towers are fundamentally distributed networks that are deployed where people live, work, and travel. In contrast, datacentres—particularly hyperscale facilities—are centralised by design. Even edge datacenters, though more distributed than hyperscale, are still limited in scope compared to telecom towers.
This difference defines everything from site management to revenue models. Comparing the operational and strategic dimensions of towers and data centers, it is clear that towercos understand how to manage thousands of distributed sites—a capability that becomes valuable in the edge computing space.
Power Consumption
Perhaps the most eye-opening contrast is in power requirements. A typical telecom tower consumes around 5 kW on average. An edge datacentre, by comparison, might consume 5 MW or more, and hyperscale centres exceed 100 MW. McKean shared that this was the steepest learning curve for him and his team: understanding the sheer complexity and cost of power provisioning.
Towercos entering the datacentre space must prioritise energy strategy, working closely with internal experts and external utilities and suppliers to ensure viability.
Risk Profiles
Tower networks distribute risk geographically and a failure at one site affects only that location. Datacentres, being centralised, concentrate risk. A single point of failure could jeopardise the entire operation. This leads to having higher security, compliance, and disaster recovery standards for datacentre operators.
Capital Expenditure
Both asset classes are capital intensive. Towers require large upfront investments, often for acquiring portfolios. Datacentres similarly demand massive initial capital deployment, especially for hyperscale builds. But while structurally different, the financial commitment and long-term ROI models are aligned.
Market Growth and Maturity
Towers are mature, stable assets that provide steady, predictable returns, making them attractive to private equity, institutional and infrastructure investors. Datacentres, by contrast, are fast-growing, fuelled by demand for cloud computing, AI, and edge applications. They offer higher potential returns but come with greater volatility and risk.
Customer Profiles
Towercos primarily serve MNOs. The relationships are long-term and well understood. Datacentre clients include hyperscalers (AWS, Azure, Google), enterprises, and government agencies. These customers speak a different language, with distinct procurement, compliance, and technical needs.
McKean emphasised that towercos must either retrain staff or recruit experienced datacentre professionals to bridge this cultural and technical gap.
Security and Compliance
Towers require physical security and real estate compliance. Data centres face additional layers: cybersecurity, data sovereignty, and rigorous regulatory scrutiny. This complexity demands new governance and risk management strategies.
Service Model Differentiation
Towers offer infrastructure and related engineering services. Datacentres may involve everything from simple colocation to full-stack services, including compute, storage, and GPU-as-a-service. The level of service commitment and customer engagement is therefore deeper and more varied in the datacentre model.
Shared Strategic Ground: Where Towers and Datacenters Align
Despite their differences, towers and datacenters share several core attributes:
Time and Capacity Sales: Both sell infrastructure over time and at scale.
Contract Models: Long-term contracts and predictable revenue appeal to similar investor classes.
Investor Convergence: Pension and sovereign wealth funds increasingly invest in both towers and data centers.
Competitive Neutrality: Towercos thrive by being neutral hosts; the same principle applies in datacenters, particularly if they avoid managing customer applications.
Emerging Models: Where Towercos Can Win
Edge Computing: Towercos are uniquely positioned to lead in edge deployments. Many of them manage macro and micro sites/ small cells &DAS. Hyperscalers excel in centralised operations but struggle with managing thousands of micro-sites.
A few are experimenting with edge deployments in venues like stadiums—spaces with existing power, security, and downtime capacity.
Repurposing Exchanges and Infrastructure: BT is investigating in retrofitting its 5,500 telecom exchanges to host edge compute capabilities. These sites are distributed, underutilised, and ideally placed to meet low-latency AI inference demand in urban and suburban areas.
Sustainability as Strategy: With datacenters consuming massive power, BT and others have emphasied sustainability. Liquid cooling, renewable integration, and energy optimisation are no longer optional—they are strategic imperatives.
Passive-Only Datacenter Models: A more cautious initial approach for towercos could be to enter datacenters passively offering power, space, and connectivity without competing on active services. This preserves neutrality and reduces operational risk. It also aligns with the tower industry's philosophy of supporting rather than replacing customer operations.
What should towercos be careful about
Focus on operational strengths—site density, local presence, and deployment agility. Don't try to out-hyperscale the hyperscalers.
Customer alignment is key. MNOs and hyperscalers think and operate differently. Misalignment here leads to failed investments – knowing the customer is paramount
Infrastructure is fragmented. Real opportunities may lie in underserved areas or unique use cases (e.g., smart cities, autonomous vehicles).
Partner with hyperscalers or enterprises rather than competing directly. Focus on providing the real estate and utilities; let others manage the compute.
Strategic Takeaways for Towercos
“Success depends on understanding the customer and staying true to core capabilities, ” advised David McKean. Towercos have deep expertise in distributed infrastructure, risk mitigation, and neutral operations. These strengths translate well to edge computing and certain types of datacentre deployments.
However, the complexity and rapid evolution of the datacentre space require humility, partnerships, and a willingness to learn. Leveraging existing assets, collaborating with experts, and starting with a neutral, infrastructure-led model are key to success.
The convergence of towers and datacentres is no longer theoretical. As AI and edge computing demands escalate, infrastructure providers must adapt. Towercos can play a vital role—particularly in edge deployments where their distributed nature, local knowledge, and operational discipline are unmatched.
By embracing their strengths and adapting strategically, towercos can become critical enablers of the AI-powered, data-driven economy of tomorrow.
