2025年世界经济展望报告_1（英_8页_1mb

AI Expansion Drives Energy Consumption Growth

The report details that AI-related economic activities primarily occur in NAICS codes 518/519 and 5415, including specialized research labs, datacenter operators, cloud providers, and vertically integrated tech companies. Additionally, AI elements are increasingly embedded across diverse economic sectors, complicating precise categorization.

Energy demand is projected to grow significantly, especially in the US and China by 2023-2025, with the US scenario (medium demand) influenced by McKinsey. Estimates suggest substantial increases in electricity usage, with AI's energy footprint even reaching levels comparable to global IT.

The data sources include Haver Analytics and IMF-ENV models, using weighted averages based on corporate revenues. The report highlights the challenge in accounting for all AI-driven energy use due to hybrid models and the complexity of sector classification beyond simple tech categorizations.

Future scenarios incorporate renewable energy trends and policies, showing potential effects on overall consumption patterns. Various agents and scenarios depict differing levels of impact, from exogenous (TEP) shocks to technology-specific multipliers within broader sectors.

The analysis underscores AI's significant role in global energy demand, influenced by both current consumption patterns and anticipated policy shifts toward renewables, alongside challenges in accurately measuring these effects across heterogeneous industries.