Artificial Intelligence and the Environment: Is AI Solving Climate Change or Making It Worse?
AI can help fight climate change, but data centres also consume growing amounts of electricity, water, and resources. Explore whether artificial intelligence is becoming part of the climate solution—or part of the problem.
By Jay Jarwar
7/13/20265 min read
Every AI Prompt Has an Environmental Cost
Every day, billions of people use artificial intelligence to write emails, generate images, answer questions, create software, and automate work. AI appears almost magical—instant, invisible, and effortless.
But behind every AI-generated response lies a vast physical infrastructure that many people never see.
The scale of that infrastructure is growing rapidly. The International Energy Agency estimates that data centres consumed about 485 terawatt-hours of electricity worldwide in 2025. Its updated 2026 outlook projects consumption could roughly double to around 950 TWh by 2030, with AI-focused data centres growing much faster than the sector overall.
Thousands of powerful servers operate around the clock inside enormous data centres. These facilities consume huge amounts of electricity, require millions of litres of water for cooling and demand a constant supply of advanced electronic equipment.
As AI adoption accelerates worldwide, an important question is emerging:
Can artificial intelligence help save the planet while simultaneously placing greater pressure on its natural resources?
The answer is more complex than many people realize.
Related Reading: Pakistan’s Solar Revolution: A People’s Answer to the Energy Crisis
The Hidden Environmental Cost of Artificial Intelligence
Unlike traditional software, modern AI systems require enormous computing power.
Training and operating advanced AI models can require large clusters of specialised processors running for extended periods, while everyday AI use creates additional electricity demand through repeated inference in data centres.
The IEA estimates that data centres accounted for around 1.5% of global electricity consumption in 2024. Although their global share remains relatively modest, demand is highly concentrated in particular regions and is growing much faster than overall electricity consumption.
Once trained, these systems continue consuming electricity every time users generate text, images, videos or computer code.
As millions of new users adopt AI every day, global demand for computing power continues to rise rapidly.
This has led technology companies to invest billions of dollars in expanding data centres across North America, Europe, Asia and the Middle East.
While these facilities are essential for AI, they also leave a growing environmental footprint.
The Water Behind Artificial Intelligence
Perhaps the least understood environmental impact of AI is its enormous demand for water.
Powerful computer processors generate tremendous amounts of heat.
To prevent servers from overheating, many data centres rely on sophisticated cooling systems that use large quantities of freshwater.
However, there is no single reliable figure for the water consumed by an AI query or workload. Research from Lawrence Berkeley National Laboratory found that workload-level water use can vary by more than 10,000-fold depending on factors including server efficiency, cooling technology, local climate, and the water intensity of the electricity grid.
In water-stressed regions, rapid data-centre development can intensify competition for local water resources unless cooling systems and site selection are carefully managed.
Water, once viewed as unrelated to artificial intelligence, is becoming one of its most valuable resources.
Energy Consumption and Carbon Emissions
Artificial intelligence cannot function without electricity.
Data centres already consume significant amounts of global electricity, and demand is expected to grow as AI applications become more widespread.
If that electricity comes from coal, oil or natural gas, AI indirectly contributes to greenhouse gas emissions.
According to the IEA, electricity used by data centres currently produces roughly 180 million tonnes of indirect CO₂ emissions annually. That represents about 0.5% of global fuel-combustion emissions. In its base case, the IEA expects those emissions to rise to around 300 million tonnes by 2035 as data-centre demand grows.
Technology companies are increasingly investing in renewable energy, nuclear power, and energy-efficient processors to reduce their carbon footprint.
The environmental impact of AI growth will therefore depend heavily on how quickly electricity systems decarbonise, how efficient AI hardware and models become, and where new data centres are built.
Beyond Electricity: Land, Minerals, and Electronic Waste
Artificial intelligence depends upon more than software.
It requires massive buildings, thousands of servers, high-speed communication networks and specialized semiconductor chips.
Manufacturing these components consumes valuable minerals such as lithium, cobalt, copper and rare earth elements.
As hardware becomes obsolete, electronic waste continues growing.
Discarded servers, processors and electronic components contain valuable materials but also hazardous substances that require responsible recycling.
As AI infrastructure expands and specialised hardware is replaced, it will add to the broader challenge of electronic waste. UNEP notes that the exact share attributable specifically to AI remains uncertain, while only about 22% of global e-waste is currently recycled in an environmentally sound manner.
AI’s Physical Footprint Extends Beyond Data Centres
The environmental footprint of AI also includes land for data centres, transmission infrastructure, semiconductor manufacturing, and the extraction of minerals used in computing equipment. These impacts vary widely by location, which is why evaluating AI environmentally requires examining its full lifecycle rather than electricity consumption alone.
Can Artificial Intelligence Help Protect the Environment?
Despite these challenges, AI also offers remarkable environmental opportunities.
Researchers are already using AI to:
Forecast extreme weather events.
Monitor deforestation from satellite imagery.
Detect illegal fishing.
Improve precision agriculture.
Optimize electricity grids.
Reduce industrial energy consumption.
Predict floods and wildfires.
Track wildlife populations.
Improve recycling systems.
The potential climate benefit could be substantial. The IEA estimates that widespread adoption of existing AI applications in energy, transport, buildings, and industry could enable emissions reductions equivalent to roughly 5% of global energy-related emissions by 2035. However, the agency stresses that these gains are not guaranteed and could be limited by infrastructure, data, skills, regulation, and rebound effects.
Rather than replacing environmental protection, AI has the potential to become one of its most powerful tools.
The Environmental Race Inside Big Tech
Technology companies are increasingly redesigning data centres to reduce their environmental footprint. Microsoft says newer cooling designs can avoid approximately 125 million litres of water use per facility each year, while Google reported replenishing about 78% of its freshwater consumption in 2025. These efforts show that AI infrastructure can become more efficient, although company-reported sustainability figures should be viewed alongside independent measurements of the sector’s overall growth.
The Real Challenge: Responsible Artificial Intelligence
The debate should not be whether society should abandon AI.
That is neither practical nor desirable.
The real challenge is developing sustainable artificial intelligence.
Related Reading: Beyond Artificial Intelligence: AGI, Human- AI Integration and What Comes Next
Future innovation must focus on:
Energy-efficient AI models.
Renewable-powered data centres.
Water-saving cooling technologies.
Longer-lasting computer hardware.
Better recycling of electronic equipment.
Transparent reporting of environmental impacts.
Governments, technology companies and researchers all have important roles in ensuring AI grows responsibly rather than unsustainably.
Final Perspective: AI Can Help Fight Climate Change—but It Also Has an Environmental Cost
Artificial intelligence is neither inherently good nor inherently bad for the environment. Its impact depends on how much computing it requires, how that electricity is generated, how efficiently data centres use water and materials, and whether AI applications produce meaningful environmental benefits elsewhere in the economy.
The same technology that increases demand for electricity can also optimise power grids, reduce industrial energy use, improve weather forecasting and help detect environmental damage.
The key question is therefore not whether society should choose between artificial intelligence and environmental protection. It is whether AI can be developed quickly enough to deliver useful economic and environmental benefits while its own physical footprint is measured, disclosed and reduced.
Artificial intelligence may become an important tool in the fight against climate change—but it cannot substitute for clean energy, effective environmental policy or responsible resource management.
Key Takeaways
Data-centre electricity demand is growing rapidly, partly driven by AI.
AI is an important driver of data-centre growth, but data centres still account for a relatively small share of global emissions.
The water footprint of AI varies greatly depending on technology, cooling systems, electricity sources and location.
AI could help reduce emissions in sectors such as energy, transport, buildings and industry.
AI’s ultimate environmental impact will depend on clean electricity, greater efficiency, transparency and responsible deployment.
About the Author
Jay Jarwar is the founder and editor of JayJarwar Insights. He writes about artificial intelligence, technology, geopolitics, economics, public policy and emerging global trends, with a focus on explaining complex issues in clear and accessible language.
JayJarwar Insights
Ideas • Analysis • Perspectives
© 2026 JayJarwar Insights. All Rights Reserved.