Can AI Grow Green? Evidence of an Inverted-U Curve Between AI, Energy Use and Emissions
TIDE Working Paper 91, “Can AI Grow Green? Evidence of an Inverted-U Curve Between AI, Energy Use and Emissions”, examines whether increased adoption of artificial intelligence can align with climate goals. Authored by Angel Melguizo, Raúl Katz, and Juan Jung, the paper analyses data from 23 middle- and high-income countries.
The findings reveal an inverted-U pattern: early AI deployment tends to increase power consumption and CO₂ emissions, largely due to energy-intensive model training. However, once AI investment reaches approximately US $220–580 per capita, this trend reverses. Countries begin to benefit from emissions reductions and a higher share of renewable energy, thanks to more efficient infrastructure, the use of pre-trained models, and greener data centres.
The research highlights key policy implications for late adopters, showing how strategic investment can help bypass the initial emissions spike and accelerate AI’s potential as a driver of sustainable development.
TIDE Centre 