Mass adoption of text-to-image diffusion models could consume roughly 2 to 9 TWh per year, based on very rough assumptions about user behavior and hardware.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Climate Implications of Diffusion-based Generative Visual AI Systems and their Mass Adoption
Mass adoption of text-to-image diffusion models could consume roughly 2 to 9 TWh per year, based on very rough assumptions about user behavior and hardware.