Water usage of AI: It’s higher than you think

Artificial Intelligence (AI) is changing our world, from creating text to analyzing data. But it uses a lot of water, mainly for cooling data centers and producing electricity. As AI use grows, so does its water footprint.

Why AI’s Water Use Matters

AI, like chatbots or image generators, runs on data centers with thousands of servers. These get hot and need cooling, often with water. Plus, the electricity powering them can come from water-heavy sources like hydroelectric or coal plants. A 2024 study estimated U.S. data centers used 17 billion gallons of water for cooling in 2023 and 211 billion more for electricity. This could double by 2028.

Why care?

Environment: Water is limited, and heavy use hurts ecosystems.

Responsibility: AI companies should reduce their impact.

Awareness: Knowing your AI use helps you make greener choices.

How AI Uses Water

Direct Use: Data centers use water to cool servers. A big center can use as much water as a small town of 10,000 people.

Indirect Use: Electricity for AI often comes from water-intensive sources like coal or hydroelectric plants. More complex AI tasks, like training models, use more water.

Calculate Your AI Water Footprint

You contribute to AI’s water use every time you use a chatbot, create AI art, or search with AI. Here’s how to estimate it:

Track AI Use: Count daily AI tasks (e.g., 10 chatbot queries, 2 AI images).

Estimate Water Per Task:

Chatbot query: ~1 liter

AI image: ~3 liters

Calculate Total:

10 queries × 1 liter = 10 liters

2 images × 3 liters = 6 liters

Total: 16 liters/day

Yearly: 16 × 365 = 5,840 liters (~1,543 gallons)

Consider Location: Data centers in dry areas or using coal power have a bigger impact. If you know the provider of your AI tool (e.g., xAI, OpenAI), check their sustainability reports to see where their data centers are located and what energy sources they use. This can refine your estimate.

Compare this to household water use: the average person in the U.S. uses about 370 liters (100 gallons) per day for activities like drinking, cooking, and washing. Your AI usage might seem small in comparison, but it adds up across billions of users.

Reduce Your AI Water Footprint

Use AI Less: Make queries efficient to cut usage.

Pick Green Providers: Choose companies using renewable energy or efficient cooling.

Support Change: Push for water-saving tech in AI.

Offset Impact: Save water elsewhere, like shorter showers.

The Big Picture

AI’s water use is part of its broader environmental impact, including energy and carbon emissions. Better tech, like efficient cooling or renewable energy, can help. You can make a difference by using AI wisely and supporting sustainable practices.

AI’s water use is a hidden issue. By calculating your AI water footprint, you can understand your impact and make smarter choices. Small steps by users and big changes by companies can keep AI sustainable.