A United Nations-backed study has quantified the enormous energy demands of artificial intelligence, calculating that the power used could supply residential electricity for 770,000 people in Sub-Saharan Africa.
The study examined the energy footprint of training advanced AI models and the broader environmental costs of the technology’s rapid growth. Researchers also projected that the swift obsolescence of AI hardware will generate millions of metric tons of electronic waste by 2030.
The expansion of AI has been accompanied by a surge in demand for computing power, which requires vast amounts of electricity to train and run large models. Data centers supporting this work consume significant energy and rely on specialized hardware that is frequently replaced.
The comparison to residential power needs in Sub-Saharan Africa underscores the scale of AI’s consumption and the inequities in global energy access. Many communities in the region still lack reliable electricity.
The findings add to growing scrutiny of the environmental impact of AI, including its carbon emissions and contribution to e-waste. The UN study calls attention to the need for more sustainable practices as the technology continues to expand worldwide.
Created by Ayen Stabel.
Stabel is AI and can make mistakes.
Sources:
https://www.devflokers.com/blog/ai-news-june-2026-models-research-developments