![]() |
| Illustration: AI-generated infographic based on the author's article, "Will AI Cause a Water Crisis?" |
AI has dominated conversations in recent years, causing equal amounts of hope and fear amongst its supporters and detractors, respectively, owing to its ability to massively enhance productivity, even while it threatens to make human employees redundant. What has perhaps gone unnoticed is that AI may be contributing to a water crisis in a world already reeling from an environmental crisis brought about by more than a century of rapacious, largely unhindered exploitation of the earth’s scarce natural resources. Is AI going to add another disappointing chapter to this sordid saga?
Water Guzzling Data Centres are the Problem
It is data centres that make AI tick, but in the process consume a lot of water to cool down the servers that help in carrying out tasks ranging from summarising a document to creating an image. What that does to a country like India, where 342 million people out of a population of 1.4 billion people don’t have access to safe water, is something that needs to be seriously looked at. Water-hungry data centre infrastructure is expected to annually consume enough water to satisfy the water needs of 1.3 billion people at the end of the decade. In India, it is projected that data centres’ annual water consumption could cater to Mumbai’s water needs for a whole week. Apart from the above, the fact that the electricity generated for running a data centre also requires water adds to the problem.
That being stated, data centres are not the only business or industrial activity that consumes a lot of water. There have been quite a few heavy industries in the past that required the use of a lot of water. The steel, textile, chemical and paper & pulp industries are cases in point in this regard. But the problem with data centres, specifically is that these are often located in areas where water resources are already very strained. The situation gets exacerbated by global warming, which can further negatively impact the availability of water.
Can you have data centres and yet avert a water crisis?
Given that no country, including the most populous one in the world like India, can afford to lag behind with regard to AI adoption, one has to look at managing the negative fallout of such a policy with regard to pressure on scarce water resources. For instance, data centres could look at rainwater harvesting or treated wastewater for their cooling needs. They could also look at using seawater for cooling the servers, something which gives coastal cities like Visakhapatnam an advantage. No wonder that Reliance Industries has decided to set up a 1.5 GW data centre there.
Improved cooling technologies like air cooling and liquid cooling, as well closed loop systems, can also help reduce the consumption of water for cooling data centre servers. Besides, the use of more efficient newer chips that require less energy and cooling than in the past helps mitigate the situation as well.
Data centre growth- planning is the key
Data centre growth may be inevitable in an era that is already defined by AI, but allowing that to happen in an unplanned, haphazard and random manner can prove disastrous for water security across nations at a time when climate change is already negatively impacting water availability across the globe. Ensuring that data centres don’t come up in regions facing water scarcity and choosing optimal infrastructure design is of the essence here. The implementation of efficient water management practices can help ensure that the growth of data centres provides an impetus to the growth of AI in countries around the world, rather than lead to debilitating water crises.






