Artificial intelligence (AI) is redefining the global energy sector, but its rapid development raises crucial questions about its environmental impact. On Wednesday, December 4, and Thursday, December 5, 2024, Paris is hosting an international conference organized by the International Energy Agency (IEA) to analyze the implications of AI for energy and climate. This groundbreaking initiative brings together more than 200 experts, industry leaders, policymakers, and tech giants from 25 countries.
The main objective is to discuss the “two sides of the coin”: the opportunities offered by AI to optimize energy resources and the environmental challenges tied to its high electricity consumption. According to the IEA, data centers, the backbone of AI technologies, currently account for 1% of global electricity consumption. In some regions, such as Ireland and Virginia (United States), this proportion already exceeds 20%.
Optimizing Resources with AI
Despite its growing energy footprint, artificial intelligence is also seen as a powerful tool for improving energy system efficiency. AI technologies can regulate the electricity consumption of buildings, optimize the production of renewable energies, and accelerate research on electric batteries that are less dependent on scarce resources. These advancements could profoundly transform the global energy transition.
However, the rise of digital solutions increases the pressure on existing infrastructures. The geographic concentration of data centers, particularly in the United States, Europe, and Asia, creates strains on electrical grids and slows connection projects, according to the IEA.
Responding to the Climate Crisis
Tech leaders such as Google, Microsoft, and Amazon are ramping up efforts to power their data centers with decarbonized energy sources such as wind, solar, and now nuclear energy. Over five years, Google’s CO2 emissions have risen by 48%, highlighting the urgency for a transition to more sustainable energy models.
The IEA calls for enhanced cooperation between the energy, technology, and policymaking sectors to address these challenges. However, as Siddharth Singh, from the IEA’s Forecast Division, explains, the available data remains insufficient to fully anticipate AI’s impact: “We know how many electric cars are produced each year, but estimating the number of data centers under construction is still complex.”
