拆解AI背后的“国际能耗账单”

Core Insights - The energy consumption of AI is becoming a global focus, impacting energy policies and climate governance [4][6] - The concept of "green computing" is emerging as a potential solution to the energy challenges posed by AI [6][11] Energy Consumption of AI - AI data centers are referred to as "electric tigers" due to their significant energy consumption, with global data center electricity usage projected to reach 415 terawatt-hours (TWh) in 2024, accounting for 1.5% of global electricity consumption [2] - By 2030, this figure is expected to rise to approximately 945 TWh, nearing Japan's total annual electricity consumption [2] - In developed economies, the increase in data center electricity consumption is substantial, with the U.S. expected to see nearly 50% of its total electricity demand growth from data centers between 2024 and 2030 [2][3] Carbon Emissions - The International Energy Agency (IEA) forecasts that global data center carbon emissions will rise from 180 million tons in 2024 to 300 million tons by 2035, making data centers one of the fastest-growing sources of emissions [3] Green Computing Initiatives - Major economies are exploring diverse energy solutions to address the energy demands of AI. For instance, 90% of electricity used by European data centers comes from renewable sources [6] - The U.S. tech giants are increasingly purchasing renewable energy through long-term agreements and exploring nuclear energy options [6] - China is accelerating the green transformation of data centers, achieving over 80% renewable energy usage in new data centers [7][8] International Cooperation - The challenges posed by AI's energy consumption are prompting a need for global cooperation in the field of green computing, including joint research and standardization efforts [9][10] - The transition towards renewable energy in AI is expected to reshape the global energy landscape, potentially leading to significant geopolitical and economic changes [10][11]