The International Energy Agency report, Modernising grids in the age of electricity, examines how digital technologies and Artificial Intelligence (AI) can be leveraged to optimise existing electricity networks as the world transitions to electricity as its primary energy carrier. Electricity demand is currently growing at roughly twice the rate of overall energy demand. To meet expected growth by 2035, grid capacity must increase by at least 30%, or approximately 25 million kilometres of lines. Expanding infrastructure is increasingly slow and costly; therefore, making better use of existing assets is a critical priority.

Executive summary

Grid expansion is currently failing to keep pace with rising demand and energy system transformations. As of 2025, at least 1,700 GW of advanced-stage renewable projects and 600 GW of utility-scale batteries were awaiting connection. Congestion costs reached USD 12 billion in the United States and EUR 4.3 billion in the European Union in 2024. The report highlights that expansion alone is not the most cost-effective solution, as removing all congestion would require sizing the grid for peak conditions that occur only rarely. Instead, the “digital grid toolkit” – comprising digitalisation, grid-enhancing technologies (GETs), and AI – offers a way to unlock latent capacity. IEA analysis estimates that deploying dynamic ratings, topology optimisation, and advanced power-flow control could allow up to 330 GW of additional generation and demand to connect without reinforcement, avoiding approximately USD 100 billion in investment.

The need for modern grids

Grids are the backbone of modern electricity systems but are becoming increasingly complex to operate. The rise of distributed energy resources, such as rooftop solar and electric vehicles (EVs), has turned many unidirectional systems into dynamic two-way flows. This transformation requires operators to manage voltage and stability more actively. While grid expansion remains indispensable, digital and AI-enabled solutions can unlock capacity more quickly and often at a lower cost. The digital grid toolkit provides the foundation for this transition through better sensing, data connectivity, and remote control of assets.

State of play

Technical capability is currently advancing faster than deployment. Digital maturity is highly uneven, with the largest gaps found at the distribution level. A survey of network operators revealed that 60% of distribution system operators (DSOs) have no controllability over at least half of their networks. Conversely, digital twins are already delivering measurable benefits, such as a 70% reduction in the time needed to analyse reinforcement options in the United Kingdom. GETs, including dynamic line ratings (DLR), are technically mature (TRL 9) but remain geographically concentrated in Europe and North America. AI is fast becoming a productivity tool for forecasting and maintenance, with grid-related AI patenting growing by more than 500% in the five years to 2022.

Estimating and unlocking the potential

The barriers to scaling these technologies are largely structural rather than technical. Surveyed operators identified skills and organisational readiness (64%), data foundations (60%), trust-related issues (60%), and regulatory frameworks (56%) as the primary obstacles. Current regulatory models often exhibit a capital expenditure (CAPEX) bias, where operators earn returns on physical assets but treat digital software as operating costs with no profit margin. To unlock the potential, the report suggests moving towards total expenditure (TOTEX) frameworks or performance-based regulation that rewards measurable outcomes like unlocked capacity and improved reliability.

Policy recommendations

The report outlines five priority areas for action. Firstly, regulatory frameworks must be reformed to incentivise efficient decisions across all possible solutions, including digital tools. This includes shifting towards output-based regulation that links allowed revenues to measurable outcomes such as capacity unlocked. Secondly, robust and shared data foundations should be built, particularly by extending monitoring at the distribution level. Thirdly, there must be a strategic investment in skills to bridge the gap between power-system expertise and data science. Fourthly, trust and security must be established through standardised validation, certification, and “cybersecurity by design”. Finally, international co-operation is essential to pool expertise, develop shared AI foundations, and provide blended finance for emerging markets and developing economies to leapfrog traditional development paths.