Artificial intelligence is often described as a digital revolution, but AI’s expanding physical infrastructure consumes significant material resources. AI’s growth uses real-world resource extraction and consumption. It’s creating a greater global demand for materials and infrastructure.
“AI’s consumption of materials is definitely a negative thing that needs to be tapered down,” says Michael Neward (10). “People think it’s just digital, but I know they’re consuming huge amounts of water. It’ll definitely have long-term effects.”
The core of AI systems is hardware components such as graphics processing units (GPUs) and AI accelerators. These chips rely on materials such as silicon and gold, which are difficult and energy-intensive to extract. The rapid scaling of AI hardware is already straining supply chains.
According to Programs, there are currently 11,038 data centers globally, located in 174 countries, and this number is predicted to triple by 2030. Between now and 2030, companies worldwide are expected to invest nearly $7 trillion in building and upgrading data centers.
Additionally, data centers rely heavily on water for cooling to prevent servers from overheating. As AI workloads grow more complex, the demand for cooling increases, leading to greater water use. The water is often taken from local freshwater sources, especially in regions already facing limited water access or drought.
The Environmental and Energy Study Institute reports that AI data centers are increasingly tapping into freshwater resources. Large data centers can consume up to 5 million gallons a day, equivalent to the water use of a town of 10,000 to 50,000 people.
The environmental impact of this level of water consumption is enormous. Much of the water is lost to evaporation during the cooling process and cannot be easily reused. As it continues to expand, this raises concerns about sustainability and competition for water resources between infrastructure and local communities. Some companies are exploring alternative methods, such as air cooling or recycling water.
Energy consumption is another major concern created by AI infrastructure. Data centers require enormous amounts of electricity to power servers and maintain cooling systems. As AI becomes more advanced, it requires more power and increasing energy demand.
Training a single large AI model can consume as much electricity as hundreds of households use in a year. This puts pressure on power grids and increases reliance on energy production systems that many still rely on, which often use fossil fuels.
“I use AI daily, but I never really knew too much about material consumption,” says Axton Underwood (10). “I started hearing about it more recently and realized that no one is really paying attention to how much it hurts the environment.”
The impact of AI extends beyond water and energy to the materials used to build the hardware. Mining for earth elements used in chips and servers often creates environmental damage. Habitats can be destroyed when mining chemicals pollute soil and water sources.
Electronic waste is another growing concern because servers and chips have short lifespans, often replaced every few years due to technological advancements. Many of these are hazardous substances that can contaminate the environment. Recycling programs exist, but are not always efficient or implemented, so typically significant portions of materials from old hardware never get reused.
Your paragraph text by Circles by Galia