Schneider Electric released a report to explore the future power consumption of artificial intelligence data centers and called on policymakers to guide it carefully. The report analyzes the potential pressure of AI development on the power network, simulates four different future scenarios: sustainable AI, growth limits, unlimited abundance and energy crisis, predicts different growth trends of AI power consumption from 2025 to 2035, and Three key recommendations were made on how to promote sustainable AI development, including optimizing data center infrastructure, improving AI model efficiency, and establishing a sound governance framework. The report aims to guide governments and industry leaders in striking a balance between AI development and environmental and economic sustainability.
With the rapid development of generative AI, investment in high-performance and power-hungry AI infrastructure continues to increase, placing huge pressure on power networks and the environment. Schneider Electric's report provides us with power consumption forecasts under different development scenarios and puts forward practical suggestions, providing a direction for sustainable development. It is worthy of in-depth study and reference by governments, enterprises and research institutions to jointly promote the development of AI technology. Healthy and sustainable development.
Schneider Electric recently released a report on the future power consumption of artificial intelligence (AI) data centers, calling on policymakers to carefully guide the power use of AI data centers. The study, titled "Artificial Intelligence and Electricity: A System Dynamics Approach," was released after the International Energy Agency's (IEA) Global Energy and AI Conference last month. The report explores different scenarios for AI-related electricity consumption, highlighting the pressure that continued growth in AI infrastructure may place on power grids.
Picture source note: The picture is generated by AI, and the picture authorization service provider Midjourney
With the rapid rise of generative AI, a large amount of investment has poured into high-performance and power-hungry infrastructure. The report points out that existing data center infrastructure already consumes a lot of energy, and it is expected that with the popularity of AI, power demand will increase. is set to rise further, raising concerns about the strain the grid may be under and its environmental impact.
The Schneider Electric model proposes four different future scenarios: sustainable AI, limits to growth, unlimited abundance and energy crisis. Although all four scenarios predict an upward trend in electricity consumption from 2025 to 2030, the subsequent trends vary depending on the assumptions.
Under the sustainable AI scenario, Schneider predicts that electricity consumption in the AI field will increase from 100 terawatt hours (TWh) in 2025 to 785TWh by 2035. In this scenario, AI infrastructure and demand form a mutually reinforcing relationship, promoting efficient use of resources. In the growth-limited scenario, the growth of AI is restricted by power and infrastructure bottlenecks. It is expected that power consumption will reach 510TWh by 2030.
The infinite abundance scenario warns that the rapid and unrestricted development of AI systems may lead to unsustainable resource utilization. It is expected that electricity consumption will surge to 880TWh by 2030 and 1370TWh in 2035. The energy crisis scenario shows the conflict between the power demand of rapidly growing AI and other economic sectors, which may lead to economic recession and energy shortages.
In the report, Schneider Electric made three recommendations to promote sustainable AI: first, optimize the next-generation data center infrastructure, including adopting the latest cooling technology and high-performance AI hardware; second, improve the efficiency of AI models through model pruning and Quantification and other technologies; finally, establish a sound governance framework and formulate certification standards for sustainable AI practices to ensure that policy and technology complement each other. The purpose of the report is to guide government and industry leaders in promoting the growth of AI while taking into account environmental and economic sustainability.
Highlights:
Schneider Electric warns that AI data center power demand needs to be carefully guided to prevent stress on the power grid.
Forecasting four future scenarios, electricity consumption will show different growth trends from 2025 to 2035.
Three suggestions are put forward to optimize infrastructure, improve the efficiency of AI models, and establish a governance framework to promote sustainable development.
All in all, Schneider Electric's report provides an important reference for the sustainable development of AI, emphasizing the importance of paying attention to energy efficiency and environmental protection while pursuing AI progress. Only through technological innovation, policy guidance and industry cooperation can the healthy development of AI be ensured and potential energy crises avoided.