ECOLAB
“ Data centre growth is colliding with real resource limits, especially energy and water. As power demand rises, it often increases upstream water impact. At the same time, AI and high-density computing are driving much higher heat loads, which can quickly raise cooling intensity and water use if they’ re not managed well.”
The scale of the problem is only escalating. Emilio points to projections showing that by 2030, AI is expected to consume power comparable to the electricity demands of India, and water equivalent to the annual drinking water needs of the entire United States. Those are not margins for negotiation.
Yet Emilio frames this not as a ceiling on growth but as a design problem.“ The opportunity is that this doesn’ t have to slow growth,” he says.“ With smarter cooling design, realtime monitoring, and continuous optimisation, operators can decouple growth from resource use. When done right, it improves uptime, efficiency and cost performance all at once.”
Water stress is now a site selection and licence-to-operate issue The geography of data centre development has shifted faster than many operators anticipated.
Since 2022, roughly two-thirds of new capacity has been added in waterstressed regions and approximately 40 % of existing facilities already face water stress exposure. Looking further ahead, nearly a third of data centres currently under construction are in areas projected to face greater water scarcity by 2050.
2030
the year when AI is expected to consume a comparable amount of electricity to India
40 %
the portion of existing data centre facilities struggling with water stress WATCH NOW
Growing fast. Growing our impact. Growing our team.
6 ecolab. com