Lina Martin
In three years, Artificial Intelligence in the US will consume more energy than the whole of Spain
Lina Martin
In three years, Artificial Intelligence in the US will consume more energy than the whole of Spain
The exponential growth of artificial intelligence (AI) is triggering an unprecedented increase in global energy demand. In the United States, data centers dedicated to this technology could consume more electricity than entire countries such as Spain, Italy or the United Kingdom in the next three years. U.S. Department of Energy (DOE) forecasts indicate that, by 2028, the energy consumption of these centers will reach record levels, posing serious challenges for the country's electrical infrastructure. While the technology industry is looking for sustainable solutions, such as nuclear power, experts warn that the speed of data center expansion could outstrip the responsiveness of the U.S. power grid.
How much energy will AI consume in the U.S. by 2028?
According to official estimates, the energy consumption of AI data centers in the United States will reach 325 terawatt hours (TWh) in 2028. This figure represents a significant increase compared to previous years and is equivalent to more than what is consumed annually by countries such as Spain (246 TWh), the United Kingdom (287 TWh) or Italy (298 TWh). In fact, AI is expected to account for between 6.7% and 12% of total electricity consumption in the North American country.
This increase in demand is driven by the growing adoption of generative AI models, which require enormous computational capacity to operate efficiently.
Is the U.S. power grid ready to cope with this growth?
Experts warn that the U.S. power infrastructure is not prepared to absorb the rapid growth in demand from AI data centers. Currently, it is estimated that the country will need between 74 and 132 gigawatts (GW) of installed capacity to supply these facilities, which exceeds Spain's total electricity generation capacity of 125.6 GW in 2023.
The lack of proper planning could lead to strains on the power grid, and some industry voices, such as OpenAI director Sam Altman, have warned that AI could trigger an unprecedented energy crisis if drastic efficiency and supply measures are not taken.
What are the solutions put forward by the technology industry?
Faced with the growing energy challenge, leading technology companies are exploring different energy sources to ensure a stable and sustainable supply. One of the most widely considered solutions is nuclear energy, especially through small modular reactors (SMRs), which offer constant energy production with less environmental impact.
Companies such as Microsoft, Google and Amazon have already closed agreements for the development of nuclear energy projects to power their data centers. The reactivation of plants such as Three Mile Island in Pennsylvania and collaborations with startups specializing in modular reactors are some of the efforts underway to meet this energy challenge.
How will this growth affect the environment?
The rise of AI data centers poses not only energy challenges, but also environmental concerns. In addition to high power consumption, these centers require large amounts of water for processor cooling systems. In 2023, U.S. data centers used around 66 billion liters of water, and it is estimated that this figure could reach 124 billion liters by 2028.
Reliance on traditional energy sources, such as coal and natural gas, could increase carbon emissions if transition strategies to renewable or nuclear sources are not adopted. Experts urge the implementation of energy efficiency policies to mitigate the environmental impact of AI.
Author: Lina Martin
Content Manager Papernest
The authors are responsible for the choice and presentation of the facts contained in this document and for the opinions expressed therein, which are not necessarily those of Tourism and Society Think Tank and do not commit the Organization, and should not be attributed to TSTT or its members.
This site uses cookies from Google to deliver its services and to analyze traffic. Information about your use of this site is shared with Google. By using this site, you agree to its use of cookies.
