When we ask artificial intelligence a question, everything seems to happen by magic. We type a question, wait a few seconds, and receive an answer. No sheet of paper, no physical product, no transportation. It seems like a cost-free activity.
In reality, every conversation with a chatbot activates a network of enormous data centers, servers, and cooling systems that consume electricity and water.
In other words, the digital world is far from immaterial.
Understanding this is important not so much to stop using AI, but to learn how to use it more consciously.
Artificial intelligence lives in data centers
Every time we interact with a virtual assistant, our request is processed by powerful computers housed in large data centers around the world.
These facilities operate 24 hours a day and require enormous amounts of electricity, both to power the servers and to keep them at a safe temperature through sophisticated cooling systems.
According to the International Energy Agency (IEA), in 2024 data centers accounted for around 1.5% of global electricity consumption. With the continued spread of artificial intelligence, energy demand could more than double by 2030.
A typical AI-focused data center consumes as much electricity as 100,000 households, while the largest facilities currently under construction will consume 20 times as much.
A query doesn’t consume energy in the same way as a Google search
Many users assume that using ChatGPT or other AI tools has a similar impact to carrying out a regular online search. In reality, this is not the case.
A query to a language model generally requires greater computing power than a traditional web search.
The IEA explains that generative AI increases the workload of data centers because the system has to process enormous amounts of data in real time, rather than simply retrieving information that has already been indexed.
As the number of users grows, this translates into greater energy requirements for digital infrastructure.
It’s not just about energy: water is also needed
One of the lesser-known aspects concerns server cooling. To prevent overheating, many data centers use large quantities of water.
While the carbon footprint generated by artificial intelligence has attracted increasing public attention, AI’s water withdrawal and consumption are still relatively little discussed.
A study conducted by the University of California, Riverside, together with the University of Texas at Arlington estimates that a conversation with a chatbot may indirectly require around half a liter of water, taking into account the cooling needed for dozens of requests. The study also estimates that training the GPT-3 language model in Microsoft’s state-of-the-art U.S. data centers can consume 700,000 liters of clean freshwater. Globally, AI-related water demand is expected to reach between 4.2 and 6.6 billion cubic meters in 2027—six times Denmark’s annual water withdrawal or half that of the United Kingdom.
The AI boom is also changing investment
The expansion of artificial intelligence is affecting more than just the technology sector.
According to McKinsey, hundreds of billions of dollars in investment will be needed in the coming years to build new data centers, power grids, and energy infrastructure capable of meeting the growing demand for computing capacity.
This means that AI has become an economic phenomenon that is reshaping the energy, construction, and even raw materials markets.
So, should we stop using AI?
The answer is “not necessarily.”
Artificial intelligence can improve productivity, support scientific research, facilitate learning, and make businesses and public administrations more efficient.
The point is not to give up the technology, but to understand its hidden costs.
Just as we learn not to waste water or electricity at home, we can learn to use digital technologies more responsibly: avoiding unnecessary requests, making the most of human intelligence, and remembering that every online activity has a real environmental impact that should not be underestimated.
Digital technology has an environmental footprint
When we hear about sustainability, we immediately think of plastic, cars, or airplanes. Streaming, the cloud, and artificial intelligence rarely come to mind.
Yet our time online relies on huge infrastructures that consume real resources.
According to the European Commission, the digital transformation must go hand in hand with the green transition, improving the energy efficiency of digital infrastructure and promoting the use of renewable energy sources so that innovation does not lead to increased emissions.
The challenge is not to choose between innovation and sustainability, but to make them grow together.
Technology is invisible. Resources are not.
Artificial intelligence gives us the impression that everything happens in a virtual space.
In reality, as we have seen, behind every answer there are computers, energy, water, power grids, and people who design and manage these infrastructures.
Understanding what happens behind the scenes helps us take a broader perspective: every innovation has a cost, and true innovation also means making that cost increasingly sustainable.
Using AI consciously means not only making the most of its opportunities, but also understanding the value of the resources that make it possible.
Discover more content on innovation, sustainability, and financial education on the Museum of Saving blog. Take part in our events, workshops, and educational programs designed to help young people, adults, families, and schools build a solid foundation for the future. Visit www.museodelrisparmio.it.