Thirsty AI: How Artificial Intelligence Threatens Water Security—and How We Can Reverse This Future. OPINION. August 20, 2025

Source: Canva Archive

Thirsty AI: How Artificial Intelligence Threatens Water Security—and How We Can Reverse This Future. OPINION. 08/20/2025

The Paradox of Innovation

Artificial intelligence is seen as the engine of the next economic revolution.

But behind the promise of smarter, more productive algorithms lies an invisible cost: water.

While 2.2 billion people still lack basic sanitation, the data centers that support AI consume enormous volumes of this resource—often in regions already vulnerable to drought.

In the US, 40% of data centers operate in areas of high water stress. A single center can use up to 5 million gallons of water per day, the equivalent of the consumption of a city of tens of thousands of people.

A 2023 study showed that training GPT-3 alone consumed 700,000 liters of water—enough to produce 370 electric cars or 320 pairs of jeans. And this is just one example. With the rise of generative AI, the sector is projected to demand up to 6.6 billion cubic meters of water per year by 2030—more than half of what the entire United Kingdom uses annually.

When algorithms drain communities

In Brazil, this contradiction is already looming.

Giants like ByteDance (TikTok) have set up data centers in Caucaia, Ceará, a region marked by a historic water crisis. There, up to 80% of the water used in cooling systems is lost through evaporation, while local communities struggle to ensure a regular supply.

This scenario is no exception. In New Mexico (USA), farms face severe shortages while data centers consume millions of liters of drinking water. In Phoenix, Arizona, residents are protesting the installation of new units precisely because aquifers are already at critical risk.

The issue is simple: we are prioritizing the thirst of algorithms over the thirst of people.

Sustainable AI is possible.

The contrast is stark when we consider the existing solutions.

One example comes from Brazil: in 2025, a national AI system was developed, hosted in data centers powered entirely by renewable energy and carbon neutrality, without increasing the water footprint (InforChannel, 2025).

In Switzerland, IBM Research's Aquasar has already demonstrated the possibility of reusing heat from servers to heat entire buildings, drastically reducing waste. Other initiatives are focusing on air cooling, closed circuits, and even the use of non-potable or recycled water instead of potable water.

The problem, therefore, is not technical. It is political and economic: most Big Tech companies still opt for the cheapest and fastest model, even if it means competing with vulnerable communities for the planet's most vital resource.

What's at Stake

If nothing changes, the rise of AI could accelerate social and environmental crises:

● Agriculture at risk: less water available for irrigation in already fragile areas.

● Social conflicts: local populations competing with global corporations.

● Energy collapse: water is also essential for hydroelectric and thermal energy; more pressure means double jeopardy.

● Climate regression: without governance, AI could widen inequalities rather than resolve them.

Possible paths

To prevent the digital revolution from becoming a digital drought, we need:

1. Transparency: public reporting on water use in each data center, not just its carbon footprint.

2. Smart regulation: restrictions on installation in critical regions and incentives for low-water-impact technologies.

3. Water justice: prioritizing human and agricultural water supply over corporate consumption.

4. Responsible innovation: scaling solutions such as air cooling, heat recovery, and the use of non-potable sources.

Conclusion:

Artificial intelligence can transform the world—but not at the expense of our water.

The future doesn't have to be one of thirsty algorithms draining vulnerable communities.

The Brazilian example of AI developed with 100% renewable and carbon-neutral data centers shows that the path exists.

The challenge is to ensure that this is the rule, not the exception.

Because without water, technology is impossible.

Anna Luisa Beserra

SDGs: SDG 6, SDG 7, SDG 11, SDG 12