The next phase of digital growth is increasingly being shaped by something far less virtual: access to reliable electricity. Artificial intelligence, cloud computing, data centres, advanced manufacturing and electrified transport are all adding new demand to power systems that were largely built for a slower-moving economy. The result is a shift in the digital investment conversation. Computing capacity still matters, but the ability to connect that capacity to power at the right location, price and timetable is becoming just as important. IEA - Electricity 2026
This changes the way companies, investors and policymakers need to think about technology-led expansion. The International Energy Agency has described a new “Age of Electricity” in which demand is rising across industry, buildings, transport and digital infrastructure. Its 2026 electricity analysis also identifies grids as a growing bottleneck, with connection queues and long project lead times increasingly influencing where new demand can be served. The implication is straightforward: digital ambition can move at software speed, while the physical systems supporting it cannot. IEA - Grids, Electricity 2026
Digital demand is colliding with physical limits
Data centres are a clear illustration of the mismatch. A new facility can move from planning to operation in a few years, yet transmission lines, substations and other grid infrastructure may require much longer periods for permitting, procurement and construction. The IEA’s work on energy demand from AI projects rapid growth in data-centre electricity consumption through 2030, driven particularly by accelerated computing. Even if data centres remain a minority share of global electricity demand, their growth can be highly concentrated in individual regions, creating local constraints that are economically significant. IEA - Electricity 2026
The practical challenge is not simply generating enough electricity at a national level. New digital facilities need capacity at specific nodes of the system and often require high standards of reliability. That places pressure on local networks, transformers, interconnection processes and backup arrangements. A market may appear to have ample generation in aggregate while still being unable to connect a large new customer on the timetable that customer requires.
Grid connection is becoming a strategic variable
For technology companies, the availability of grid connections is starting to resemble other scarce inputs such as land, specialist labour and advanced chips. Location decisions can therefore become more infrastructure-sensitive. Sites with access to transmission capacity, predictable permitting and a credible pipeline of new generation may command a strategic advantage over locations where connection queues are long or upgrade requirements are uncertain. IEA - Grids, Electricity 2026
This dynamic can also reshape regional competition. Governments seeking to attract AI, cloud and advanced manufacturing investment may find that tax incentives and industrial policy are less effective if the underlying network cannot support new loads. In that sense, electricity infrastructure is becoming part of the investment proposition for the digital economy rather than a background utility service.
Capital is moving toward the bottleneck
As constraints become more visible, capital allocation can move toward the parts of the system that limit expansion. The IEA has estimated that global grid spending is running well below investment in generation, despite rising demand and the need to connect both new supply and new users. This gap helps explain why transmission equipment, cables, transformers, substations, storage and grid-management technologies are becoming more prominent in infrastructure strategies. IEA - Energy demand from AI
The investment opportunity, however, is not simply a case of spending more. Grid projects have long asset lives, regulated returns in many markets and exposure to permitting, supply-chain and cost risks. The most valuable projects may be those that unlock constrained capacity, improve reliability or reduce the time needed to connect new demand. Investors therefore need to distinguish between broad infrastructure expansion and assets with a clear role in relieving a binding bottleneck.
Existing networks may have more value than they appear to
New construction is only one response. Because grids are designed around peak conditions, parts of the system can have unused capacity at other times. Digital monitoring, dynamic line ratings, reconductoring, flexible demand and storage can help make more use of existing assets. The IEA has highlighted a range of grid-enhancing technologies that can release additional hosting capacity more quickly than conventional network expansion. IEA - World Energy Investment 2025
This creates a broader lesson for the digital economy: the solution to infrastructure scarcity may come partly from better utilisation rather than only from new supply. Companies that can shift computing loads, improve energy efficiency or coordinate demand with system conditions may gain more flexibility over where and when they expand. Utilities, in turn, may need new commercial and operational tools to manage large, flexible customers.
Power procurement is becoming part of technology strategy
For large digital companies, energy procurement is no longer only a sustainability or facilities function. It can affect deployment schedules, operating costs and the feasibility of entire projects. Long-term contracts, on-site generation, storage, demand response and direct partnerships with utilities are increasingly relevant because they can reduce uncertainty around both supply and connection. IEA - Building the Future Transmission Grid
The same logic applies to companies that depend indirectly on cloud and AI infrastructure. If power scarcity pushes up data-centre costs or slows capacity additions, those effects can flow into cloud pricing, service availability and the economics of AI deployment. The physical cost base beneath digital services therefore deserves more attention in technology budgets and long-range planning.
The supply chain is part of the constraint
Even where capital is available, infrastructure cannot be expanded instantly. The IEA’s work on transmission has highlighted tight supply chains and longer lead times for key equipment. Transformers and cables are not interchangeable commodities that can always be sourced at short notice; they can involve specialised manufacturing capacity, technical standards and long order books. IEA - Grid investments
This introduces a different kind of execution risk. A project can have financing, approvals and demand but still face delays because critical components are unavailable. Companies and infrastructure operators may respond by placing orders earlier, standardising designs, building supplier relationships and carrying more inventory. Such measures can raise near-term costs but lower the risk of missing strategic windows for capacity expansion.
A new test for growth forecasts
Analysts have traditionally modelled digital growth through demand, market share, computing costs and customer adoption. Increasingly, infrastructure availability belongs in that framework. A forecast for rapid AI or data-centre expansion may be economically plausible yet operationally unrealistic if the required power, network capacity or equipment cannot be delivered on time.
This does not mean digital growth is likely to stop. It means the pathway may be less smooth and more geographically uneven than headline demand forecasts imply. Regions with strong infrastructure may absorb a disproportionate share of investment, while constrained markets may face higher costs or slower deployment. The value of infrastructure readiness could therefore become more visible in corporate valuations and regional economic performance.
The digital economy is becoming more physical
The broader trend is a convergence of technology and infrastructure. AI may be software, but delivering it at scale requires land, electricity, cooling, networks, semiconductors and construction. As demand rises, the economics of these inputs can become as important as the models and applications running on top of them.
That convergence changes what counts as technology strategy. The winners may not simply be the organisations with the most advanced software. They may also be those that can secure scarce physical capacity, use it efficiently and plan around infrastructure lead times. In an economy increasingly shaped by digital services, one of the most important competitive questions may be surprisingly basic: where will the power come from?
Key questions
Why is electricity becoming a constraint on digital growth?
AI, cloud computing and data centres are increasing power demand faster than some local grids can expand, while grid projects often have longer lead times than digital facilities.
Does this mean data centres will overwhelm global power systems?
Not necessarily. Their global share remains limited, but demand is concentrated in particular regions, so local network constraints can still be significant.
What could reduce the bottleneck?
New grid investment, faster permitting, better equipment supply chains, energy efficiency, storage and technologies that use existing networks more effectively can all help.
References
IEA - World Energy Investment 2025