Chatbots and Water Scarcity: The Hidden Cost of AI
It was November 2022 when the world witnessed a paradigm shift in the technological realm with the introduction of OpenAI’s ChatGPT, or Chat Generative Pre-trained Transformer. Since then, almost every field associated with technology has experienced disruption. Alongside ChatGPT, other prominent chatbots such as Google’s Gemini, Microsoft’s Copilot, and the Chinese giant DeepSeek have also become prevalent in the industry. Their user base is growing by leaps and bounds with time.
These “helpers” of humanity are hailed as a step forward, however, humans’ growing reliance on them may have disastrous consequences, far beyond the understanding of a common individual. A befitting quote that applies here is: “Everything comes with a price.” While its origin is uncertain, the meaning resonates clearly: humans have gained easy access to knowledge that was once difficult to obtain. With just one click, yes, only one click, ChatGPT can provide millions of results based on prompts and keywords. It sounds great, doesn’t it? But every convenience comes with a cost. What price does ChatGPT demand? Strangely enough, it is humanity’s scarcest and most vital resource, water.
OpenAI, the creator of ChatGPT, has headquarters in San Francisco, London, and Dublin. These offices house the basic infrastructure to run AI servers. Chatbots are not powered by a single computer, rather they rely on massive, heavy-duty cloud servers. OpenAI uses a network of computers equipped with high-end GPUs (graphics processing units) to run large language models (LLMs) such as ChatGPT. This powerful machinery ensures smooth operations but requires enormous amounts of energy. That energy comes from electricity, and electricity generation, whether from hydropower or grid systems, demands water.
Furthermore, the servers used by AI companies must be kept cool to function. The primary means of cooling is, once again, water. In simple terms, ChatGPT and other chatbots run on servers, these servers consume large amounts of power, and ultimately, this power and cooling process depends heavily on water. AI, therefore, poses a significant and under-discussed threat to water resources.
In 2025, the United Nations Educational, Scientific and Cultural Organization (UNESCO) and UN-Water published The World Water Development Report 2025, highlighting the worsening global water scarcity. The report warned that nearly half of the world’s population will soon face severe water stress. UNICEF further noted that four billion people already suffer from water shortages for at least one month each year, and around 700 million could be displaced by 2030 due to water scarcity. By 2040, it estimates that one in four children will live in water-scarce regions. Similarly, the Food and Agriculture Organization (FAO) of the UN cautioned that urbanization, industrialization, and population growth will exacerbate this crisis. Clearly, water is humanity’s crown jewel, and it is under threat.
Ironically, AI, once celebrated as the solution to human dilemmas, is now becoming a looming danger. The Food and Water Watch Organization published an article warning that by 2028, the U.S. AI industry alone may consume as much electricity as 28 million households, magnifying the water crisis further. After detailed research, the organization estimated that by 2030, the U.S. AI sector could consume around 300 terawatt hours (TWh) of electricity annually. To keep its servers cool, the U.S. AI industry may use up to 720 billion gallons of water per year, enough to supply nearly 18.5 million households.
Adding to this, former President Trump announced a 500 billion dollar investment to modernize and accelerate AI development, which could further increase water use. To put this into perspective, one ChatGPT query consumes nearly ten times more energy than a Google search. These figures are only the tip of the iceberg, revealing the scale of damage AI could cause in the near future.
Humanity now stands on the brink of a drought-driven world where resources, especially water, may become triggers for wars. While technology is indispensable, ways to mitigate its harmful effects must be devised. First, public awareness is essential, people need to understand the environmental costs of AI overuse. Second, world leaders and international organizations must step up. Developed nations should devise strategies to conserve water and invest in large-scale recycling systems, as their stronger economies can fund such initiatives. Third, responsibility lies with computer engineers and tech innovators to create environment-friendly infrastructure and reduce the unquenchable thirst for power in AI systems.
Ultimately, infrastructure development is crucial to addressing the environmental challenges posed by AI. Without urgent measures, humanity risks paying a catastrophic price for the convenience of artificial intelligence.
Disclaimer: The views and opinions expressed in this article are solely those of the author and do not necessarily reflect the editorial stance, policies, or official position of The Spine Times.
