AI now consumes more water than the global bottled water industry: New study reveals

Recent research has shed light on the significant environmental impact of artificial intelligence, revealing that its water usage now surpasses that of the entire global bottled water industry. The study, led by Dutch researcher Alex de Vries-Gao, titled 'The carbon and water footprints of data centres and what this could mean for artificial intelligence,' highlights alarming trends in how AI technology consumes vital water resources. According to the findings, the annual water withdrawal by the AI sector has escalated to more than 450 billion liters, which is the amount of bottled water consumed yearly worldwide. This surge is driven by the immense cooling requirements of the data centers that support large language models and generative AI technologies. With advanced processors from companies like Nvidia generating excessive heat during operation, these facilities depend on extensive water resources to maintain optimal temperatures. The report points out that many data centers utilize evaporative cooling systems or liquid cooling loops, drawing from local water supplies or already stressed aquifers. This has prompted a growing concern about a 'transparency gap' in the way major tech companies manage their water consumption. While firms like Microsoft, Google, and Meta have made commitments to achieve 'water positive' status by 2030, the actual water usage rates have surged in tandem with AI advancements. For instance, a single interaction with a chatbot can result in an estimated water consumption of 500 milliliters due to evaporation at the data center level, influenced by local server conditions and temperatures. As many of these centers are located in water-scarce regions, tensions with local communities and agricultural interests are on the rise. The study underscores the need for a fundamental shift toward more efficient hardware and closed-loop cooling systems that minimize evaporation. As international regulatory bodies begin to formulate environmental reporting standards for AI, experts assert that the focus must expand beyond carbon footprints to include the water usage of digital technologies. The escalating water footprint of AI poses a significant challenge to the sustainability of its ongoing development.