As AI technology expands, data centres are facing scrutiny over their substantial water consumption. In 2025, these facilities consumed an estimated 222 billion litres of water for cooling, a figure projected to nearly triple by 2030 without intervention. Tech giants like Nvidia claim their new closed-loop cooling systems could significantly reduce this demand, but the trade-off may be increased energy use.
Nvidia’s innovative approach involves circulating warmer water directly through servers, potentially lowering water needs but requiring more electricity for cooling. This shift highlights a critical dilemma: as companies strive to meet public pressure for sustainability, they may inadvertently escalate energy consumption, complicating their environmental impact.
Moreover, the indirect water usage associated with electricity generation and chip manufacturing often surpasses the direct consumption of data centres. This means that while tech firms can tout reductions in water use, the overall ecological footprint remains significant, raising questions about the effectiveness of their claims.
As public opposition to water-intensive data centres grows, the industry must balance technological advancements with environmental responsibilities. The challenge lies in innovating solutions that genuinely reduce both water and energy consumption, ensuring sustainable growth in the AI sector.
Source: Euronews
