Chart of the Day

AI's Electricity Bill Becomes an Asset-Owner Problem

The AI trade is moving from chips and software into power markets, utilities, transmission, cooling, data centres and infrastructure capital.

Line chart of global data-centre electricity demand rising from 415 TWh in 2024 to about 945 TWh by 2030, roughly 15% annual growth, just under 3% of global electricity demand.

The IEA's base case projects global data-centre electricity consumption will more than double to around 945 TWh by 2030, growing roughly 15% per year from 2024 — reaching just under 3% of global electricity demand. The AI trade is migrating from chips and software into power, grid, cooling and land.

Commercial insight

The investable opportunity may migrate from public AI equities into physical bottlenecks: power generation, grid upgrades, substations, cooling, land, water, transmission, fibre and real estate. The asset managers that can package AI infrastructure for pensions, sovereign funds and insurers may become the next major capital-formation winners.

Allocator insight

Universal owners may already be heavily exposed to AI through public equities. The second-order question is whether they also own the infrastructure required to make the AI economy physically possible — and whether that infrastructure is being priced as a technology bet or as long-duration real assets.


Source: International Energy Agency, Energy and AI (2025).

Subscribe

The morning briefing for the people who allocate long-horizon capital.

Research, charts, video and podcast analysis for the institutions investing at the scale of the world.

Five minutes, five days a week. Free.