UN calculates nation-sized environmental footprints for AI and data centers
3 hours ago
- #UN energy report
- #AI environmental impact
- #data centers sustainability
- UN report highlights that data centers' environmental footprint rivals major countries, with electricity usage surpassing all but 10 nations in 2023.
- AI drives significant growth in data centers, expected to increase from 20% to 40% of their energy consumption by 2030, doubling water and energy use in four years.
- Energy use by data centers in 2023 produced about 208 million tons of CO2, comparable to Argentina, and consumed 1.2 trillion gallons of water.
- By 2030, data centers are projected to account for nearly 3% of global electricity use, ranking sixth-highest if they were a country, with CO2 emissions reaching 440 million tons.
- The report emphasizes the hidden environmental costs of AI, noting its infrastructure and hardware have real impacts despite appearing 'clean' without visible pollution.
- Reducing word use in AI queries by 30% can cut energy consumption by 25%, saving electricity equivalent to what 700,000 people in Africa use annually.
- AI efficiency gains may lead to increased overall energy use due to higher demand, and reliance on renewable energy for data centers can deplete clean supplies elsewhere.
- Lack of transparency from companies regarding data center and AI consumption poses challenges for managing and mitigating environmental impacts.