show index hide index
- Projections indicate that AI-related water demand could reach between 312 and 765 billion liters by 2025. For comparison, this volume rivals the annual global consumption of bottled water. The environmental impact of these digital tools is evident not only in data exchange but also in the massive use of vital resources. By 2024, AI energy consumption had already surpassed Bitcoin mining, raising numerous questions about the long-term viability of these technologies.
- Second, the electricity required to power these powerful computing systems also comes at a water cost. The power plants that supply data centers consume significant amounts of water to operate. De Vries-Gao highlights the fact that this indirect consumption is often underestimated in the sustainability reports of large technology companies.
- . This figure, which seems negligible at first glance, becomes concerning when considered on the scale of daily users. Google has acknowledged that its flagship AI, Gemini, consumes the equivalent of
- For some players, this is giving rise to alarming local tensions. In regions already affected by water stress, such as Iowa in the United States or certain areas in Europe, data centers are entering the competition for access to a resource as precious as water. The study highlights the need for greater transparency in the reports of technology companies, which often conflate the specific impact of AI with that of their other cloud services.
By 2025the artificial intelligence revolution will no longer be measured solely by the power of its algorithms, but also by its insatiable thirst forwater. A shocking study reveals that AI’swater consumption could exceed that of all the bottles used by humanity in a year. This alarming finding raises crucial questions about the environmental impact of technologies we take for granted. As tech giants battle for dominance in this sector, the need to rethink our approach becomes increasingly urgent. An alarming finding on water consumption By 2025, artificial intelligence (AI) could exceed the total bottled water consumption of all humanity. A study conducted by researcher Alex de Vries-Gao of VU Amsterdam reveals that AI’s water footprint has exploded, reaching staggering figures that constitute a serious environmental problem. As tech giants like Microsoft, Google, and Meta continue to dominate the generative AI sector, it becomes crucial to recognize the colossal impact on our water resources.
An Explosion in AI Water Demand
Projections indicate that AI-related water demand could reach between 312 and 765 billion liters by 2025. For comparison, this volume rivals the annual global consumption of bottled water. The environmental impact of these digital tools is evident not only in data exchange but also in the massive use of vital resources. By 2024, AI energy consumption had already surpassed Bitcoin mining, raising numerous questions about the long-term viability of these technologies.
The Sources of AI’s Thirst The need for such a large quantity of water for AI stems from two distinct but equally critical sources. First, the intense heat generated by the GPUs used to train and run advanced models like GPT-4 or Gemini necessitates water-cooling systems. This process results in massive evaporation, rendering this water unavailable for other uses, including agriculture and drinking water.
Second, the electricity required to power these powerful computing systems also comes at a water cost. The power plants that supply data centers consume significant amounts of water to operate. De Vries-Gao highlights the fact that this indirect consumption is often underestimated in the sustainability reports of large technology companies.
The True Cost of a Simple Interaction with AI Every request made to a chatbot has an environmental cost. According to research from the University of California, Riverside, a series of 10 to 50 questions asked of a generative AI uses approximately
50 centiliters of water
. This figure, which seems negligible at first glance, becomes concerning when considered on the scale of daily users. Google has acknowledged that its flagship AI, Gemini, consumes the equivalent of
five drops of water per request, a figure which, when multiplied by millions of users, becomes a significant environmental issue.A Growing Ethical and Political Challenge The observation of an increase in AI water consumption, exceeding 30%
For some players, this is giving rise to alarming local tensions. In regions already affected by water stress, such as Iowa in the United States or certain areas in Europe, data centers are entering the competition for access to a resource as precious as water. The study highlights the need for greater transparency in the reports of technology companies, which often conflate the specific impact of AI with that of their other cloud services.
While the promise of AI capable of providing solutions to climate problems continues to circulate, the physical reality of its resource demand raises major concerns. If innovation does not quickly move towards closed-loop cooling methods or more energy-efficient models, this digital revolution could inevitably fizzle out, lacking the water to satisfy its overly ambitious goals.
