Executive summary

Secure and affordable electricity systems that support economic growth and development depend on grid infrastructure. Electricity demand has grown at roughly twice the rate of overall energy demand over the past decade, driven by the electrification of end uses, by cooling and industrial expansion, and in some markets by data centres. Meeting its expected growth by 2035 requires grid capacity to increase by at least 30%, equivalent to adding or replacing 25 million kilometres of lines. Three quarters of the increase to 2035 is expected in emerging market and developing economies, where expanding networks will remain indispensable, while in many advanced economies demand is growing again after years of stagnation, adding pressure to networks that often also require modernisation.

Grid expansion is not keeping pace with rising electricity demand and the broader transformation of energy systems. The consequences are already visible: at least 1 700 GW of advanced-stage renewable projects and 600 GW of utility-scale batteries were awaiting connection in 2025, while congestion cost USD 12 billion in the United States and EUR 4.3 billion in the European Union in 2024. Permitting procedures, supply-chain bottlenecks, labour shortages and rising equipment costs all constrain delivery.

Expansion alone would not be the most cost-effective solution even if networks could be built faster. Removing all congestion would mean sizing the grid for peak conditions at all times, and because wind, solar, and many new loads run at full output only part of the time, such a network would be used well below capacity for most of the year. A degree of congestion is therefore efficient, and since network costs make up a substantial share of consumer bills, how well existing capacity is used directly affects affordability.

While grids are expanding, they are also becoming more complex to operate. Distributed resources, bidirectional flows and increasingly variable and concentrated sources of supply and demand are making network conditions change more quickly and less predictably. Operators increasingly need to actively manage voltage and stability alongside power flows and thermal limits, often with incomplete visibility.

Existing networks could be used more efficiently and carry more than they do today, both by changing how risk is managed in operation and by getting more from grid assets. Operators maintain margins below the network’s physical limits to cover risks they cannot fully observe, so improved information can allow those margins to be narrowed safely. Getting greater output from existing assets means both carrying more power through equipment already in service and keeping that equipment available more of the time, by reducing outages and shortening restoration and replacement times. This report focuses on these network-based solutions, where the underlying physics and many operating practices are common across countries. Digitalisation, grid-enhancing technologies and AI together provide a toolkit for identifying and using this capacity safely and efficiently.

IEA analysis estimates that dynamic ratings, topology optimisation and advanced power-flow control could allow up to 330 GW of additional generation, storage and demand to connect to existing networks without reinforcement. This is roughly equivalent to the combined installed electricity generation capacity of France and Korea. Connecting this volume through network expansion instead would require around USD 100 billion of investment. The estimate covers the three technologies most widely deployed in transmission networks: dynamic ratings, which adjust how much power lines and transformers can safely carry according to actual weather conditions; topology optimisation, which reconfigures the network to route power around congested lines; and advanced power-flow control, which pushes power onto underused paths. Other digital and grid-enhancing technologies would add further capacity. 

These solutions complement grid expansion rather than replacing it. Most can be deployed in months rather than years, relieving constraints while new infrastructure is being built. They can also be added to networks already in service, while asset replacement provides the lowest-cost opportunity to embed much of the required digital capability. Where conditions are favourable they can defer or avoid individual reinforcements, but the capacity they release is bounded by the physical network and varies with weather and system conditions, so they cannot deliver the step change in capability that demand growth requires.

Digitalisation and grid-enhancing technologies are already delivering measurable benefits in real networks. Reported results include USD 64 million a year in avoided congestion costs from dynamic ratings in the United States, more than 2 GW of additional transfer capacity from advanced power-flow control in Great Britain, and a 70% reduction in the time needed to analyse reinforcement options using a digital twin, a detailed virtual model of the network.

However, proven technologies do not yet move consistently from individual projects to standard practice. Deployment remains highly uneven, and technologies proven at one site frequently stall between pilot deployment and standard practice, since each deployment has to be justified afresh in the absence of established methods for valuing benefits or transferring experience between projects.

AI adds new applications and can make the wider toolkit easier to target, operate and scale. Network operators have used AI-based methods for many years for forecasting, inspection, maintenance and planning, and recent advances allow models to process much larger and more diverse operational, weather and imagery datasets. AI can help operators identify where unused capacity exists, forecast when it can be used safely and test whether it would remain if equipment failed or other system conditions changed, but adoption varies by function. Planning and asset-related applications are relatively mature, while safety-critical operational control remains constrained by explainability, auditability and accountability.

The main obstacles to wider deployment are institutional and organisational, rather than technological performance. Among surveyed network operators, 64% identified skills and organisational readiness as a major barrier, 60% availability and quality of data, 60% trust-related issues and 56% regulatory frameworks. These four conditions apply across digitalisation, grid-enhancing technologies and AI, though AI places the highest demands on data quality, specialist skills, testing and assurance, and governance. Each condition points to a front for action, and international co-operation is a fifth opportunity that cuts across all of them, allowing much of what each country needs to be built once rather than developed repeatedly. Priorities will differ across countries according to market structure, development priorities and the age and digital maturity of infrastructure, but the underlying challenges are widely shared.

Regulatory frameworks should allow operators to propose whichever solution best addresses a given network need and efficiently recover its costs. Operators often face clearer pathways to deploy new physical infrastructure than digital or operational alternatives addressing the same network need. Solutions include funding rules that do not favour capital spending on new equipment over digital or operational solutions, incentives linked to measurable outcomes such as the unlocking of capacity and avoision of congestion, and credible methods for valuing system-wide benefits that would support more efficient investment decisions. Regulatory sandboxes, which let operators trial new approaches under temporary exemptions from normal rules, and innovation funding offer a route to building evidence for tools that have yet to prove themselves.

Better data and greater visibility are most urgently needed at distribution level, where operators have limited information and the need for data is increasing fastest, by rooftop solar, electric vehicles and heat pumps being connected in growing numbers. Many distribution network operators lack visibility over large parts of their network, and information that does exist frequently sits in incompatible formats across organisational silos. Improved data quality and visibility, and interoperable data models allowing information to move between systems and trusted mechanisms for sharing data would support both digitalisation and AI. Meanwhile, shared data infrastructure and testing environments can also lower costs and spread learning.

Utilities and regulators need to build combined power-system and digital expertise, and to adapt the processes around it. Professionals who combine the two disciplines remain scarce and the grid workforce is ageing, with the number of workers aged 55 and over rising by nearly 50% between 2015 and 2024, compared with 20% for those under 55. Training should be accompanied by changes to workflows and operational processes so that proven tools move beyond pilots and become part of routine operations.

Assurance and security should be built in from the outset and calibrated to the criticality of each application. 70% of surveyed operators use AI widely for maintenance, but only 23% use AI in real-time operations. This is a reflection of the explainability, auditability and accountability requirements when AI informs safety-critical decisions, because operators remain accountable. Common approaches to validation, certification and human oversight would support wider adoption, with cybersecurity treated as a core design requirement and particular attention to where operational data are stored, who controls the models used in critical functions and how strategic dependencies are managed.

International co-operation can reduce costs, accelerate deployment and widen access to proven solutions, helping countries make more efficient use of existing infrastructure, improve affordability and accelerate electrification in a secure way. Deployment of grid-enhancing technologies has so far been concentrated in Europe and North America. Sharing expertise, testing facilities and approaches to validation and certification helps countries build trust in new technologies and manage strategic dependencies as grids come to rely more heavily on software, data and AI. For emerging market and developing economies, proven regulatory approaches, shared technical resources and targeted grants and below-market loans can help utilities deploy digital solutions earlier, build visibility and controllability through asset replacement cycles and access skills that are difficult to develop alone.