The invisible water consumed by AI questions the transparency of sustainability reports

The exponential growth of artificial intelligence is drawing attention to an impact that has so far largely gone unnoticed: its water footprint. Although imperceptible to the user, every interaction with AI systems involves a water consumption linked to cooling data centers, the true 'physical core' of this technology. According to recent estimates, a single query can require small amounts of water, but its widespread use — multiplied by millions of daily interactions — acquires significant scale.

Projections are clear: international organizations like UNESCO estimate that water consumption linked to AI could reach between 4.2 and 6.6 trillion litres in the coming years, even exceeding the annual consumption of some countries. All this in a context of increasing water scarcity, especially in regions like southern Europe, where Spain is already facing high levels of water stress and a high risk of desertification.

In this context, the CNMV has reiterated that companies must disclose their water footprint when relevant, considering both the intensity of AI use and geographic factors and risk exposure. This change marks a turning point: sustainability is shifting from being solely focused on carbon footprint to progressively incorporating water impact as a key variable for management and compliance.

Read the full article here.