As artificial intelligence continues to evolve at a staggering pace, a new concern has emerged: the substantial amount of water required to power this groundbreaking technology. Recent discussions highlight the environmental implications of AI usage, particularly regarding the water consumption of data centres. This article delves into the complexities of AI’s water footprint and the differing perspectives surrounding its impact on our precious resources.
The Thirsty Tech: How AI Consumes Water
AI technologies, such as ChatGPT and Google Gemini, rely heavily on data centres, which require significant amounts of water for cooling and operational purposes. While estimates of water consumption vary widely, the figures are alarming. For instance, ChatGPT is said to require less than 1/15 of a teaspoon of water per query, according to OpenAI’s CEO, Sam Altman. In contrast, research from the University of California in 2023 reveals that ChatGPT may consume around 500ml of water for every 10 to 50 medium-length responses.
The UK government’s Digital Sustainability Alliance predicts that by 2027, AI could increase global water usage from 1.1 billion to 6.6 billion cubic metres—a staggering projection that raises red flags about the sustainability of our water supplies.
Why Does AI Need So Much Water?
The primary reason for AI’s substantial water consumption lies in the infrastructure of data centres. These facilities not only house powerful servers but also require water for cooling systems to prevent overheating. A mid-sized data centre can use as much water as a small town, while larger ones may consume up to 5 million gallons daily, equivalent to the water usage of a city with a population of 50,000.
Moreover, the manufacturing processes that create the hardware necessary for AI also contribute to water consumption, further exacerbating the issue. The Lincoln Institute of Conventional Policy underscores the criticality of this situation, warning that AI’s reliance on local water supplies could place immense pressure on communities already facing water scarcity.
Growing Concerns: The Environmental Impact
The environmental implications of AI’s water usage are becoming increasingly concerning. Experts warn that as AI technology expands, the demand for water will likely escalate, particularly in areas already grappling with limited resources. Almost 68% of data centres are located near protected biodiversity areas, leading to fears that AI’s water consumption could jeopardise vital ecosystems.
Members of the Government Digital Sustainability Alliance have highlighted the potential for demand for freshwater to exceed supply by 40% by the end of the decade. With over half of the world’s data centres situated in river basins at high risk of water pollution, the pressure on clean water supplies is intensifying.
Is the Concern Overstated?
While many experts assert that AI’s water consumption poses a significant crisis, some argue that fears may be exaggerated. Andy Masley, director of Effective Altruism DC, suggests that the water used for a single individual’s interactions with AI is minimal compared to everyday water use. He notes that the water footprint of hundreds of thousands of ChatGPT prompts would still be less than the water required to produce a single pair of jeans.
Masley claims that an individual would need to submit over 1,000 prompts a day to increase their daily water footprint by merely 1%. He argues that generating prompts from home could even result in a lower water footprint compared to typical electricity usage.
Sam Gilbert, a researcher at the University of Cambridge, echoes this sentiment, emphasising that the focus should be on the broader environmental impact of data centres rather than individual consumption. He calls for increased transparency from companies regarding the true environmental costs of their operations.
The Future of AI and Water Consumption
As the demand for AI technology surges, the challenge lies in balancing innovation with sustainability. Shaolei Ren, an engineering professor at the University of California, warns that many data centres experience peak water usage during hot summer months, straining public water systems. Thames Water has already cautioned that data centres may face restrictions during heatwaves, highlighting the urgent need for a sustainable approach.
However, there is a glimmer of hope. AI technologies can also contribute to water conservation. For instance, AI has been deployed to detect leaks and improve energy efficiency in water distribution systems. In 2024, a water company in Surrey began using AI solutions to minimise leaks, and the World Economic Forum predicts that AI-enabled water management could potentially reduce water use by 15% in the United States.
Google claims that its data centres are among the most efficient globally, utilising air-cooling techniques to limit water consumption and pledging to replenish 120% of the water it consumes.
Why it Matters
The conversation surrounding AI’s water consumption is crucial not only for environmental sustainability but also for the future of technology itself. As AI continues to permeate our daily lives, understanding its impact on essential resources is paramount. The balance between harnessing the benefits of AI and protecting our planet’s finite water supply will define the future of both industries and ecosystems. It’s a challenge that requires immediate attention, collaboration, and innovative solutions to ensure that technological progress does not come at the expense of our most vital resource.
