Reality Signal
Artwork 03 · The Physical Cost of AI
If the digital world keeps growing, how much physical space will its future need?
The digital world depends on physical infrastructure. Projections for the next few years point to growing electricity demand and a larger land footprint associated with supplying that electricity.
The research
Electricity
485 TWh 950 TWh
2025 estimate 2030 projection
Annual electricity consumption of all data centres worldwide
The International Energy Agency projects global data-centre electricity consumption to rise from 485 TWh in 2025 to 950 TWh in 2030. That is nearly twice the electricity in five years. Electricity consumption from AI-focused data centres is projected to triple over the same period.
Land
6,900 km² 14,500 km²
2025 estimate 2030 projection
Land footprint associated with global data-centre electricity demand
A United Nations University report calculates a land footprint of around 6,900 km² associated with data-centre electricity demand in 2025, rising to around 14,500 km² in its 2030 scenario. This concerns the land associated with supplying electricity, including energy infrastructure, reservoirs and fuel extraction.
What the research does not show
Both comparisons concern all data centres, not AI alone. The land figure is not the area occupied by data-centre buildings or land newly developed each year.
The figures come from separate analyses. The UN land calculation uses a different electricity-demand projection from the updated IEA figures shown above.
These are estimates and scenarios, not a measurement of one image or your own AI use. They do not establish the net environmental impact after possible savings from AI elsewhere.
Open question
Which choices could change this projected footprint?
Sources
- International Energy AgencyKey Questions on Energy and AI, 2026. Executive summary.Source for electricity-demand estimates and projections.Read the analysis (opens in a new tab)
- Aczel et al., United Nations University Institute for Water, Environment and HealthEnvironmental Cost of AI’s Energy Use: Carbon, Water and Land Footprints, 2026. Pages 26 and 32–33.Source for land-footprint calculations.Read the report (opens in a new tab)
