Artificial Intelligence: We generally consider Artificial Intelligence to be a thing of the digital world, but there is a side to it which is directly related to the environment. A new research of the year 2025 has claimed that the water used to run AI systems may now be more than the total consumption of bottled water worldwide. Not only this, the carbon emissions related to AI this year are said to be equal to that of a big city like New York. If these figures are close to the truth, then the increasing demand for AI comes as a big warning for the environment.
How does AI consume water?
Nowadays you might have heard that AI is drinking water or AI has finished the water. In fact, AI itself does not drink water, but the machines and data centers that run it use a lot of water. This is the reason why this issue is under discussion.
When ChatGPT, Google or any AI system works, huge data centers run behind it. There are thousands of servers installed in these data centers which keep getting heated continuously. To keep them cool, a cooling system is used and lakhs of liters of water is spent in this cooling. At many places this water goes directly into the water cooling system, while at other places the consumption of water used during power generation is also added to it.
It also takes a lot of energy to train an AI model. When big companies create new AI models, the servers run at full capacity for weeks or months. During this period, not only the consumption of electricity increases but a huge amount of water is also spent in generating the same electricity and keeping the system cool. According to an estimate, there are some AI tasks in which even answering a question leads to indirect consumption of a few milliliters of water.
What does research say about the environmental impact of AI?
This peer-reviewed study has been published under the name “The carbon and water footprints of data centers and what this could mean for artificial intelligence”. This research was led by Dutch researcher Alex de Vries-Gao. The focus of the study was on those data centers on which AI systems run and which cause huge consumption of energy and water. The research also acknowledged that it is not easy to extract accurate figures because companies do not show AI and non-AI workload separately in their environmental reports.
How were the estimates made?
Due to lack of direct data, researchers adopted a different approach. They analyzed environmental reports, average emissions data and water consumption data related to data centers of big tech companies like Google, Meta, Amazon. Based on this, it was estimated how much electricity and water is required to handle the AI workload.
As much carbon as New York, more consumption than bottled water
According to the study, in the year 2025, carbon emissions from AI systems alone could be between 32.6 million to 79.7 million tons of CO₂. This quantity is considered equal to the annual carbon footprint of a big metropolis like New York City. The figures of water consumption are even more shocking. Data centers connected to AI can use about 312 to 764 billion liters of water every year, which is more than the bottled water used in a year in the entire world. It is clear from this that AI is becoming a big issue not only of energy but also of water security.
No training, daily use is becoming a big reason
An important thing has come to light in the research that the biggest impact on the environment is not caused by training AI models but by their daily use i.e. inference. When users ask questions, create images or videos and digital assistants work day and night, there is a heavy load on data centers. Due to crores of requests, electricity and water consumption increases rapidly.
Better technology, but the effect is not decreasing
Surprisingly, despite efforts to make data centers more energy-efficient, the overall environmental impact is not being reduced. The reason is clear, the use of AI is increasing so rapidly that all efforts for improvement fall short in front of it. In simple words, technology is getting better but its use is increasing even faster.
It is no longer right to consider AI as just software
Two big things emerge from this study. First, AI should no longer be seen as just software. Just as environmental regulations apply to telecom, aviation and heavy industries, the same level of monitoring should be done on the AI industry. The second important thing is transparency.
The research says that if companies do not clearly share data on energy and water consumption related to AI workloads, it will be difficult to make the right policy. Without clear information, neither conservation nor proper planning for the future is possible. This study clearly indicates that the future of AI will not only be decided by being smart but it will also have to be seen how sustainable it proves to be for the earth.
Also read:
Is your smartphone also fake? Know how to identify the real one
