The artificial intelligence boom is creating an unprecedented surge in electricity demand from data centres, with major implications for power markets, grid infrastructure and energy prices globally. As AI models grow larger and more widespread, the energy footprint of AI computing infrastructure is becoming a significant factor in electricity market outlooks for 2026 and beyond.

How Much Power Does AI Actually Use?

Training large-scale AI models is extraordinarily energy-intensive. Training a single large language model can consume millions of kilowatt-hours of electricity — equivalent to the annual energy use of hundreds of households. And unlike model training, which is a one-time process, inference (running the model to answer queries) happens billions of times per day across deployed AI products, creating a continuous and growing electricity load.

The International Energy Agency estimated in its most recent data centre report that global data centre electricity consumption could double by 2026 compared to 2022 levels, driven primarily by AI workloads. In the US, data centres are already responsible for approximately 2% of total electricity consumption, and some forecasts suggest this could reach 4–6% by the end of the decade.

Geographic Concentration Creates Grid Pressure

AI data centre expansion is highly geographically concentrated, creating acute pressure on specific regional electricity grids. Northern Virginia — the data centre capital of the world — faces growing grid capacity constraints, with utility Dominion Energy announcing significant grid investment plans to accommodate the load. Ireland, which hosts major hyperscaler data centres due to its connectivity and tax environment, has also faced acute grid pressure, prompting the country’s electricity regulator to temporarily pause new data centre connections in the Dublin area.

The Renewable Energy Dimension

Major technology companies — including Google, Microsoft, Meta and Amazon — have made public commitments to power their data centres with 100% renewable energy. This is driving significant corporate power purchase agreements (PPAs) for wind and solar energy, which is accelerating renewable energy sector investments in markets where these companies operate.

However, the intermittent nature of renewables means that data centres — which require 24/7 reliable power — cannot operate on renewable energy alone without battery storage technology or grid backup, making the 100% renewable claim more complex in practice than it appears on paper.

What It Means for Energy Prices

The AI-driven surge in electricity demand is a positive for energy prices in markets where data centres are concentrated, as it creates incremental demand for power and reduces the risk of oversupply. For investors, utilities and grid operators serving major data centre markets represent an interesting thematic opportunity as AI infrastructure buildout continues.