The rapid expansion of artificial intelligence (AI) applications is reshaping the energy landscape and putting the U.S. economy at risk due to electricity supply constraints. CoBank has warned that the U.S. is just 12 to 24 months away from experiencing insufficient electricity to power the growing number of data centers, which currently account for about 4% of the nation's energy demand. This demand is expected to rise significantly as AI technologies proliferate [8b1f55d5].
Sundar Pichai, CEO of Google/Alphabet, has emphasized that AI's impact is more profound than fire or electricity, with U.S. electricity demand predicted to rise by 2.8% annually until 2030. This increase is notably higher than the 0.4% growth seen since 2000, with AI, electric vehicles, and other emerging technologies projected to account for one-third of this demand increase [cabfb622].
Fitch Ratings has reported that capital expenditure (capex) growth for North American companies is expected to accelerate in 2025, with capex margins rising to 8% from 7.5% in 2024. This increase is attributed to the adoption of AI and the transition to a lower-carbon economy, with electricity demand projected to grow by 2.0% to 2.5% annually until 2030, driven by data center expansion [67b290f2].
In addition to energy concerns, JPMorgan Chase & Co. has reported that the rising demand for AI is also straining U.S. water supplies. Large data centers, essential for AI operations, can consume up to 5 million gallons of water daily for cooling purposes. This demand is exacerbated by climate change, which is already contributing to water shortages across the country [b7bd2c21].
Energy companies are responding to the surge in demand by accelerating the construction of natural gas power plants, with projections indicating that natural gas will constitute up to 60% of new electricity generation capacity by 2032. Over 200 gas installations are currently in development, expected to provide approximately 86 gigawatts of electricity [c1175b68]. However, nearly 80% of U.S. consumers are feeling the strain of high energy costs, prompting utilities to request over $18.13 billion in rate increases in 2023. Major utilities have seen their delivery costs rise by 65% from 2010 to 2020 [8b1f55d5].
The electricity usage by AI data centers is projected to increase tenfold by 2030, raising concerns about the sustainability of energy resources. A single inquiry made through ChatGPT consumes ten times more energy than a standard Google search, highlighting the significant energy demands of generative AI applications [8b1f55d5].
In Georgia, energy needs surged by over 38% in just one year, further emphasizing the urgent need for federal action to address infrastructure challenges and ensure a stable energy supply [8b1f55d5]. Northern Virginia, noted for having the most data centers globally, exemplifies the strain on local power grids, which have been graded a C- by the American Society of Civil Engineers [cabfb622]. Microsoft and Google, two leaders in AI development, are grappling with the environmental implications of their energy consumption, as their data centers have contributed to a notable rise in greenhouse gas emissions [7821f5ed]. Both companies are committed to reducing their emissions and transitioning to renewable energy sources, but the rapid growth of AI technologies complicates these efforts [d5ae60dd].
The ongoing developments in the energy sector underscore the need for innovative solutions to mitigate the environmental impacts associated with AI's increasing energy demands and to ensure that economic growth is not stifled by electricity supply constraints. Furthermore, water scarcity, which could slow global GDP by up to 6%, is becoming increasingly recognized by investors as a financial risk, highlighting the interconnectedness of energy, water, and economic stability [b7bd2c21].