In the current landscape of artificial intelligence (AI), energy shortages are emerging as a significant obstacle to technological advancement in the United States. As numerous data centers await activation, there are concerns about the nation’s ability to develop adequate power infrastructure to meet the increasing demand. Historically, between 2010 and 2020, data centers were able to balance rising electricity needs with improvements in energy efficiency. However, since then, electricity demand has surged, attributed to the high volume of queries processed by popular AI models. This uptick in demand has not been matched by efficiency advancements, resulting in rising electricity costs for consumers in regions impacted by data center operations.
To support the continued advancement of AI without imposing excessive energy costs on the general population, the U.S. could potentially benefit from examining energy strategies employed in other countries. For instance, in 2024, China added 429 gigawatts (GW) of new power generation capacity, significantly surpassing the net capacity increase in the U.S. during the same period. While coal remains a substantial part of China’s energy mix, its reliance on coal is decreasing, as the nation invests heavily in solar, wind, nuclear, and natural gas technologies.
In contrast, the U.S. appears to be focusing on revitalizing coal power, which is characterized by high operational costs and diminished reliability due to aging plants. As a result, coal-fired power plants now operate at only 42% of capacity, down from 61% in 2014. If the U.S. does not adapt its energy policy, it risks becoming a consumer rather than a leader in both energy generation and AI development. Current trends suggest that China may earn greater revenue from renewable energy exports than the U.S. achieves from traditional oil and gas exports.
Source: https://www.technologyreview.com/2025/11/09/1126805/the-state-of-ai-energy-is-king-and-the-us-is-falling-behind/