UN report warns: The effect of artificial intelligence in the physical world is growing
Artificial intelligence is often perceived as an invisible technology that works on our screens. A question asks, creates a visual or can produce long texts in a few seconds. However, there is a huge physical infrastructure behind this digital experience.
The new report published in June 2026 of the United Nations University of Water, Environment and Health Institute (UNU-INWEH) is not only through energy consumption; water usage, land need and electronic waste production.
According to the report, global data centers consumed about 448 teravat-hour electricity in 2025. This amount will leave behind the total annual energy usage of some countries. The water consumption of data centers for the same year has reached the level to fill about 1.8 million Olympic swimming pools.
The researchers emphasize that artificial intelligence requires more electricity, water and infrastructure. That’s why technology is not only software; data centers, energy generation, cooling systems, chip production and electronic wastes should be evaluated as a physical system.
What is the growth of artificial intelligence on environmental resources? And what can be the long-term effects of this growth?
[
[The physical world of the bank
Yapay intelligence is often seen only as software. Yet every query works a large physical infrastructure behind every visual production and every data process.
Data centers are located in the center of this infrastructure. These facilities with thousands of servers are constantly working to train artificial models and process billions of users every day. The servers need electricity to work. Cooling systems are activated to prevent overheating. These systems consume considerable amount of water and energy.
According to the United Nations report, global data centers used about 448 teravat-hour electricity in 2025. If data centers were a country, the world’s largest electrical consumer would be included.
However, energy consumption is only part of the table. The production of electricity used for the work of data centers also needs additional resources. Energy production processes bring effects such as water usage, land need and greenhouse gas emissions.
One of the important points taken into account the report is that the environmental impact of the artificial intelligence is not limited only with data centers. Electronic equipment that completes the critical minerals, transmission infrastructures, cooling systems and usage life required for chip production is part of this chain. [
Therefore, researchers evaluate artificial intelligence not only as a digital service, but a large industrial system dependent on natural resources.
Water, energy, land and e-waste: Four footprints of artificial intelligence[
The combined Nations report examines the environmental effects of artificial intelligence under four basic titles: energy consumption, water use, land need and electronic waste.
] Energy
Data centers consumed about 448 teravat-hour electricity in 2025. According to the report, about five of which was caused by artificial intelligence applications.
This amount is about 1,3 billion people living in Substrate Africa, which can meet the need of housing electricity for years. If data centers were a country, the world’s largest electrical consumer would be included.
The researchers expect to grow further in the coming years of this demand with the rapid increase of the use of artificial intelligence.
Water
One of the least spoken environmental impacts of data centers water consumption. [email protected] Servers produce significant amount of heat when working. Therefore, water-based cooling systems are used in many facilities. In addition, electricity production processes that provide energy to data centers also consume water indirectly.
According to the report, data centers used about 1.8 million Olympic swimming pools in 2025. [
If the current growth trend continues, it is estimated that the annual water consumption of artificial intelligence in 2030 can reach 9,3 trillion liters. [
Argin Use
] Digital technologies are often considered as physical space-free systems. However, data centers, energy infrastructures, transmission networks and production facilities require significant land use.
According to the report, the area covered by the infrastructure that supports the electrical demand of data centers reaches about 7,000 kilometers. This corresponds to an area of about 4.5 times in London metropolis area.
With the growth of artificial intelligence applications, the need for new data centers and energy investments is increasing.
Electronic waste
Another topic that stands out in the radio is electronic wastes. [
Artificial intelligence systems require high performance processors, graphics cards and special chips. Due to the rapid development of technology, these equipment can be replaced by new ones in short time. [
The researchers predict that the amount of artificial intelligence-borne electronic waste can reach about 2.5 million tons until the end of the decade.
[
This figure means that the electronic equipment equivalent to the weight of hundreds of large buildings every year remains unusable.
All this data shows that artificial intelligence is not only a digital transformation; but also a physical conversion associated with energy, water and natural resources. [
Why important?
] [Yapay zekâ is one of the fastest growing technologies in recent years Home. According to the United Nations report, the global artificial intelligence market is expected to approach 5 trillion dollars by growing about 25 times in the next decade.
This growth means that the demand for data centers will increase rapidly. Mean more processing power, more servers, more energy and more cooling needs. [
However, the main issue according to researchers is not only the size of the source consumption. The problem is that the large part of these effects is invisible.
While most users print a text, when creating a visual or producing a video, there is no need for energy, water and infrastructure behind it. Artificial intelligence is often perceived as an independent technology from the physical world.
Kaveh Madani, Director of the United Nations University of Water, Environment and Health Institute, states that the discussions in the public are still focused on software size. Yet artificial intelligence; data centers, electrical production, cooling systems, transmission networks, chips, minerals, land usage and direct connected infrastructure system with water resources.
Another issue highlighted in the report is a lack of transparency. The researchers argue that data centers and AI providers should share more data on environmental impacts. Thus, both policy makers and users can make more informed decisions.
Artificial intelligence can provide efficiency in many areas, accelerate scientific research and contribute to climate solutions. However, the report reminds that the environmental costs of this conversion should be taken into account. [
[
New question of digital conversion
Yapay intelligence will be integrated into more areas of life in the coming years. It is expected to increase the use of medical education from scientific research to public services.
Therefore, the discussion is no longer going to grow and grow artificial intelligence. The question, which sources and how to manage this growth.
The new report of the United Nations reminds that artificial intelligence is not only a digital technology. It is based on a wide physical infrastructure that extends from data centers to electrical networks, water sources to electronic waste.
This indicates that environmental policies with technology policies will be increasingly cut. The transparency of data centers’ energy resources, water usage, electronic waste management and supply chains will be among the important titles of the next period.
The way to reduce the environmental effects of artificial intelligence according to experts does not slow technological development. Establishing more efficient data centers, increasing renewable energy usage, developing cooling systems that reduce water consumption and managing electronic wastes with a cyclic economy approach is among the important parts of the solution.
Artificial intelligence has the potential to contribute to sustainability targets in many areas from climate models to energy efficiency. However, in order to realize this potential, it is necessary to make the technology’s own environmental footprint appear.
The digital world has physical limits. This is the most important message that the new report revealed: the future of artificial intelligence will be shaped not only with algorithms, but also with energy, water and natural resources. [
[
sources
• United Nations University Institute for Water, Environment and Health (UNU-INWEH), The Environmental Footprint of Artificial Intelligence, 2026
• Earth. Org, 9.3 Trillion Liters of Water: UN Report Exposes Unfathomable Footprint of Data Centers as AI Booms, 5 June 2026
Home International Energy Agency (IEA), Energy and AI
Home International Telecommunication Union (ITU), Digital Sustainability studies
The new report published in June 2026 of the United Nations University of Water, Environment and Health Institute (UNU-INWEH) is not only through energy consumption; water usage, land need and electronic waste production.
According to the report, global data centers consumed about 448 teravat-hour electricity in 2025. This amount will leave behind the total annual energy usage of some countries. The water consumption of data centers for the same year has reached the level to fill about 1.8 million Olympic swimming pools.
The researchers emphasize that artificial intelligence requires more electricity, water and infrastructure. That’s why technology is not only software; data centers, energy generation, cooling systems, chip production and electronic wastes should be evaluated as a physical system.
What is the growth of artificial intelligence on environmental resources? And what can be the long-term effects of this growth?
[
[The physical world of the bank
Yapay intelligence is often seen only as software. Yet every query works a large physical infrastructure behind every visual production and every data process.
Data centers are located in the center of this infrastructure. These facilities with thousands of servers are constantly working to train artificial models and process billions of users every day. The servers need electricity to work. Cooling systems are activated to prevent overheating. These systems consume considerable amount of water and energy.
According to the United Nations report, global data centers used about 448 teravat-hour electricity in 2025. If data centers were a country, the world’s largest electrical consumer would be included.
However, energy consumption is only part of the table. The production of electricity used for the work of data centers also needs additional resources. Energy production processes bring effects such as water usage, land need and greenhouse gas emissions.
One of the important points taken into account the report is that the environmental impact of the artificial intelligence is not limited only with data centers. Electronic equipment that completes the critical minerals, transmission infrastructures, cooling systems and usage life required for chip production is part of this chain. [
Therefore, researchers evaluate artificial intelligence not only as a digital service, but a large industrial system dependent on natural resources.
Water, energy, land and e-waste: Four footprints of artificial intelligence[
The combined Nations report examines the environmental effects of artificial intelligence under four basic titles: energy consumption, water use, land need and electronic waste.
] Energy
Data centers consumed about 448 teravat-hour electricity in 2025. According to the report, about five of which was caused by artificial intelligence applications.
This amount is about 1,3 billion people living in Substrate Africa, which can meet the need of housing electricity for years. If data centers were a country, the world’s largest electrical consumer would be included.
The researchers expect to grow further in the coming years of this demand with the rapid increase of the use of artificial intelligence.
Water
One of the least spoken environmental impacts of data centers water consumption. [email protected] Servers produce significant amount of heat when working. Therefore, water-based cooling systems are used in many facilities. In addition, electricity production processes that provide energy to data centers also consume water indirectly.
According to the report, data centers used about 1.8 million Olympic swimming pools in 2025. [
If the current growth trend continues, it is estimated that the annual water consumption of artificial intelligence in 2030 can reach 9,3 trillion liters. [
Argin Use
] Digital technologies are often considered as physical space-free systems. However, data centers, energy infrastructures, transmission networks and production facilities require significant land use.
According to the report, the area covered by the infrastructure that supports the electrical demand of data centers reaches about 7,000 kilometers. This corresponds to an area of about 4.5 times in London metropolis area.
With the growth of artificial intelligence applications, the need for new data centers and energy investments is increasing.
Electronic waste
Another topic that stands out in the radio is electronic wastes. [
Artificial intelligence systems require high performance processors, graphics cards and special chips. Due to the rapid development of technology, these equipment can be replaced by new ones in short time. [
The researchers predict that the amount of artificial intelligence-borne electronic waste can reach about 2.5 million tons until the end of the decade.
[
This figure means that the electronic equipment equivalent to the weight of hundreds of large buildings every year remains unusable.
All this data shows that artificial intelligence is not only a digital transformation; but also a physical conversion associated with energy, water and natural resources. [
Why important?
] [Yapay zekâ is one of the fastest growing technologies in recent years Home. According to the United Nations report, the global artificial intelligence market is expected to approach 5 trillion dollars by growing about 25 times in the next decade.
This growth means that the demand for data centers will increase rapidly. Mean more processing power, more servers, more energy and more cooling needs. [
However, the main issue according to researchers is not only the size of the source consumption. The problem is that the large part of these effects is invisible.
While most users print a text, when creating a visual or producing a video, there is no need for energy, water and infrastructure behind it. Artificial intelligence is often perceived as an independent technology from the physical world.
Kaveh Madani, Director of the United Nations University of Water, Environment and Health Institute, states that the discussions in the public are still focused on software size. Yet artificial intelligence; data centers, electrical production, cooling systems, transmission networks, chips, minerals, land usage and direct connected infrastructure system with water resources.
Another issue highlighted in the report is a lack of transparency. The researchers argue that data centers and AI providers should share more data on environmental impacts. Thus, both policy makers and users can make more informed decisions.
Artificial intelligence can provide efficiency in many areas, accelerate scientific research and contribute to climate solutions. However, the report reminds that the environmental costs of this conversion should be taken into account. [
[
New question of digital conversion
Yapay intelligence will be integrated into more areas of life in the coming years. It is expected to increase the use of medical education from scientific research to public services.
Therefore, the discussion is no longer going to grow and grow artificial intelligence. The question, which sources and how to manage this growth.
The new report of the United Nations reminds that artificial intelligence is not only a digital technology. It is based on a wide physical infrastructure that extends from data centers to electrical networks, water sources to electronic waste.
This indicates that environmental policies with technology policies will be increasingly cut. The transparency of data centers’ energy resources, water usage, electronic waste management and supply chains will be among the important titles of the next period.
The way to reduce the environmental effects of artificial intelligence according to experts does not slow technological development. Establishing more efficient data centers, increasing renewable energy usage, developing cooling systems that reduce water consumption and managing electronic wastes with a cyclic economy approach is among the important parts of the solution.
Artificial intelligence has the potential to contribute to sustainability targets in many areas from climate models to energy efficiency. However, in order to realize this potential, it is necessary to make the technology’s own environmental footprint appear.
The digital world has physical limits. This is the most important message that the new report revealed: the future of artificial intelligence will be shaped not only with algorithms, but also with energy, water and natural resources. [
[
sources
• United Nations University Institute for Water, Environment and Health (UNU-INWEH), The Environmental Footprint of Artificial Intelligence, 2026
• Earth. Org, 9.3 Trillion Liters of Water: UN Report Exposes Unfathomable Footprint of Data Centers as AI Booms, 5 June 2026
Home International Energy Agency (IEA), Energy and AI
Home International Telecommunication Union (ITU), Digital Sustainability studies
Source: News