AI growth in EU could lead to higher emissions abroad - German environment agency
Clean Energy Wire
Artificial intelligence (AI) calculations could be increasingly performed in regions with cheap, non-renewable energy to meet growing demand, which could lead to higher overall emissions, found a report by the German Environmental Agency (UBA). The rapid growth of AI data centres outside the EU could mean companies shift their activities abroad to cut costs or avoid stringent climate regulations. Projected growth is currently concentrated in the U.S. and China, two countries where climate regulations are laxer than in the EU. AI-related energy consumption in data centres will increase to around 300 terawatt hours by 2028, which corresponds to around one percent of global electricity consumption.
The extent of carbon leakage, the phenomenon where emission reductions in one country or region lead to increased emissions elsewhere, is currently moderate, as the focus is on building capacity rather than cutting costs, the report found. Existing infrastructure and a stable energy supply appear to play the biggest role in the choice of data centre location. However, the report highlighted that carbon leakage could become a bigger issue in the future, as priorities in the industry might change. Cross-border data flows could also increase imported emissions. To avoid the weakening of European climate targets, the report recommended that renewable energy capacity be expanded in areas with strong AI growth and that AI companies comply with their own climate targets.
Energy giant RWE recently called upon the German government to prioritise the development of AI. The previous government implemented a scheme in 2023 to reduce carbon leakage associated with higher fuel costs from the country’s emission trading scheme.
