The Cloud Has a Body: What AI’s Climate Story Leaves Out is the first report published by Effective Thoughts, produced in collaboration with SMCSE Solutions and Wisdom Inn for Skill Enhancement (WISE). It contributes to the global discourse on artificial intelligence and its environmental consequences.
Key findings
- The footprint is concentrated. There are more than 11,000 data centers worldwide, and over 4,400 of them are in the United States.
- Electricity demand is set to more than double. Global data center demand could reach roughly 945 TWh by 2030. That’s more than Japan uses today.
- Much of the water use is invisible. Roughly 60% of a typical data center’s water footprint comes from generating its electricity, not from on-site cooling.
- Rules exist, but on paper. Only 36% of EU data centers required to report their energy and water data actually did.
- Efficiency is not the same as reduction. AI is getting more efficient per task, but total demand keeps rising, because people run more computation when it gets cheaper.
The full report is free to share, in full and with attribution, for non-commercial purposes under a CC BY-NC-ND 4.0 license.
