A structural shift in infrastructure investing

Data infrastructure—data centers, fiber networks and towers—has become a core area of infrastructure investing, supported by strong underlying demand from cloud computing, mobile usage and artificial intelligence (AI).

These trends are well understood. What is less appreciated is that connectivity, and the constraints around it, are increasingly determining returns. As a result, outcomes are no longer defined simply by demand alone, but by where infrastructure can be delivered and which assets can be successfully developed and expanded over time.

Data infrastructure operates as an integrated system, where performance depends on access to power, location and network connectivity. Constraints in any of these inputs can limit the ability to deliver capacity, even where demand is strong.

What was once a broad-based buildout opportunity has become more selective. In this environment, scale of capital and execution capability are critical.

For example, mobile broadband coverage now reaches 96% of the global population, yet a usage gap of more than three billion people—nearly 40% of the world—persists (Figure 1). This suggests that there is significant scope for increased data usage and the need for the infrastructure to deliver it. 

Connectivity, digital maturity and opportunity

Digital infrastructure spans a set of interconnected assets that enable the generation, transmission and storage of data. These include data centers, which provide compute and storage capacity; fiber networks, which carry high-capacity traffic; telecom towers and small cells, which support wireless connectivity; and satellite systems, which extend coverage to underserved regions. 

Together, these assets form an integrated system, with each layer dependent on the others to deliver connectivity at scale. 

Infrastructure development remains uneven across regions, reflecting different stages of digital maturity (Figure 2). In some markets, the priority remains expanding basic coverage. In others, rapid adoption is driving network densification and increasing demand for fiber and interconnection. More developed markets, while highly penetrated, are increasingly constrained by limits on power, land and network capacity.

Accelerating dynamics

AI is increasing the scale and intensity of activity across data infrastructure, while intensifying existing limitations—particularly around power and location. The outlook for data center electricity demand is highly uncertain, driven by factors including efficiency improvements, AI uptake and potential energy sector bottlenecks.

Global data center electricity demand is projected to exceed 1,000 terawatt-hours (TWh) by 2030 under high-growth scenarios, reflecting the rapid expansion of AI-driven workloads (Figure 3). This represents a step change in required infrastructure. Meeting this level of demand requires a substantial expansion of power generation, grid capacity and interconnection—each of which is subject to long development timelines and structural constraints.

Defining success

In this environment, the most successful investments share a consistent set of characteristics.

They combine access to scarce inputs such as power, land and interconnection with business models that provide revenue visibility and the ability to expand over time. Long-term agreements underpin cash flows, while size and utilization allow infrastructure to be densified, extended or replicated, improving utilization as demand grows.

These attributes are critical in a market where outcomes today are shaped by physical and structural limitations. Assets that control key inputs and can expand within constrained environments tend to benefit from durable competitive advantages.

Delivering these investments requires more than capital. It depends on the ability to secure power early, navigate permitting, coordinate development and align capacity with long-term demand.⁴ The mismatch between infrastructure demand and delivery timelines is particularly evident in power systems (Figure 4). Infrastructure projects can take approximately 20%  longer than planned and exceed budgets by up to 80%, underscoring the importance of disciplined execution.

Opportunity in a constrained market

Data infrastructure represents a core allocation within institutional portfolios, combining defensive characteristics with exposure to long-term digitalization trends.

The opportunity will continue to grow, but the ability to capture it will depend on where infrastructure can be delivered and how effectively real-world constraints can be navigated.

Returns will increasingly be differentiated by access to power, connectivity and strategic locations, as well as the capital and execution capability required to bring infrastructure to market.

Not all capacity is equal. Not all locations are replicable. Not all growth translates into returns.

In this environment, successful platforms combine access to constrained inputs with contracted, visible cash flows and the ability to expand over time. These characteristics allow investments to convert structural limitations into durable, infrastructure-like returns. 

In our view, the next phase of data infrastructure investing will be defined not by who builds the most, but by who can consistently identify, secure and execute on the most constrained opportunities.