AI’s Growing Environmental Impact: A Hidden Cost of Innovation
By the end of the year, the carbon footprint of global AI systems for the whole of 2025 could equal that of New York City. At the same time, AI’s thirst for water could rival that of the world’s bottled water market, according to new estimates.
AI is everywhere and expanding at what seems like an exponential rate. The more we use it, the more data centers are needed to run the complex calculations that power AI. These facilities are not only huge guzzlers of electricity but also of water, wich is needed to cool servers and to generate electricity in power plants.
But exactly how much electricity and water they use is something of a mystery because accurate data is hard to come by. When big tech companies like Google, Amazon and Meta release their energy data they lump it all together, making it impractical to see the potential environmental cost of AI. Some companies even refuse to report the water used by power plants, claiming it is out of their control.
In a paper published in the journal Patterns, Alex de Vries-Gao, a Ph.D. candidate at Vrije Universiteit Amsterdam, and her colleagues attempt to untangle these numbers.
The carbon and water footprints of data centers and what this could mean for artificial intelligence
Data centers, the backbone of our digital world, are facing increasing scrutiny regarding their environmental impact. A new study published in Patterns highlights the meaningful carbon and water footprints associated with these facilities, especially as demand surges due to the rapid growth of artificial intelligence (AI). The research, led by Alex de Vries-Gao, provides a complete analysis of the energy and water consumption of data centers globally, and projects future impacts based on current trends.
The study reveals that data centers currently consume approximately 1% of global electricity, a figure expected to rise dramatically with the proliferation of AI applications. Training large language models (LLMs),like those powering chatbots and image generators,requires immense computational power,translating to substantial energy use. This energy consumption directly contributes to greenhouse gas emissions, particularly in regions reliant on fossil fuels for electricity generation.
Though, the environmental impact extends beyond carbon emissions. Data centers require vast amounts of water for cooling purposes. The study emphasizes that water usage varies considerably depending on the cooling technology employed.evaporative cooling, while efficient, consumes substantial water resources, posing a risk in water-stressed regions.Option cooling methods,such as liquid cooling,offer potential water savings but come with their own set of challenges,including higher upfront costs and energy requirements for pumping.
The research also points to the geographical concentration of data centers. Certain regions, such as Northern Virginia in the United States and Ireland, host a disproportionately large number of these facilities. This concentration exacerbates local environmental pressures, particularly on water resources and electricity grids. The study suggests that diversifying the location of data centers and prioritizing renewable energy sources are crucial steps towards mitigating these impacts.
Moreover, the study explores the potential for technological innovations to reduce the environmental footprint of data centers. Improvements in server efficiency, the adoption of advanced cooling technologies, and the advancement of more enduring AI algorithms are all identified as promising avenues for reducing energy and water consumption. Though, the authors caution that these innovations must be deployed rapidly and at scale to keep pace with the exponential growth of AI.
“The future sustainability of AI hinges on our ability to address the environmental challenges posed by data centers,” says Alex de Vries-Gao. “Without significant changes in energy sourcing, cooling practices, and algorithmic efficiency, the environmental cost of AI could become unsustainable.”
more facts:
alex de Vries-Gao, The carbon and water footprints of data centers and what this could mean for artificial intelligence, patterns (2025)
2025/12/18 19:14:25