AI is evolving rapidly. With tools like OpenAI’s GPT-4, Google’s Bard, and Microsoft’s Copilot, content creation, programming, and automation are becoming faster and more efficient. As AI grows more advanced, so do the ambitions of tech giants, each racing to build more powerful systems. But beneath AI’s seamless outputs lies an uncomfortable truth: its environmental cost.
How AI Is Impacting the Environment?
AI doesn’t operate in isolation—it relies on data centers, which house the computing infrastructure that powers it. Demand for these facilities has surged, with U.S. investments doubling in just two years and similar trends in China and the EU.
These centers consume vast amounts of water for cooling during construction and operation. AI infrastructure alone could soon use six times more water than Denmark, worsening global shortages that already affect a quarter of the population.
AI also requires massive energy, mostly from fossil fuels, leading to high carbon emissions. A single ChatGPT request uses 10 times the electricity of a Google Search, according to the International Energy Agency.
Another growing concern is electronic waste. AI-driven technology and the broader tech industry generate large amounts of e-waste, often containing hazardous substances like mercury and lead, posing serious environmental and health risks.
How Much Energy Do Data Centers Use?
According to the International Energy Agency’s report Electricity 2024: Analysis and Forecast to 2026, data centers, cryptocurrency mining, and AI collectively consumed 460 terawatt-hours (TWh) of electricity in 2022—nearly 2% of global electricity demand. This is expected to increase to 1,050 TWh by 2026, which is equivalent to the total electricity use of Germany.
Training AI models like GPT-4 requires a lot of computing power, which leads to significant CO₂ emissions. For instance, training GPT-3 produced emissions similar to the annual output of 100 gasoline-powered cars. But the energy demands don’t stop once the models are trained. Even when deployed, AI models continue to consume large amounts of electricity to process queries, generate responses, and refine outputs.
The increasing energy demands driven by AI are putting major tech companies’ climate goals at risk as well. Google, for instance, has set an ambitious target of achieving net-zero emissions by 2030. Since 2007, the company has been carbon neutral, primarily through purchasing carbon offsets—funding projects that reduce CO2 emissions to balance its environmental impact. However, in 2023, Google announced that it could no longer maintain this status due to the rising energy demands from AI operations. Despite this, the company states they are still pushing for its net-zero goal in 2030.
Similarly, Microsoft had aimed to become carbon-negative within a decade, a goal that seemed within reach before the explosion of AI technology. But in 2023, the company’s emissions increased by 30%, largely driven by its growing investments in AI infrastructure. As AI continues to fuel energy consumption, these companies are being forced to reassess their ambitious climate targets.
In 2024, Elon Musk’s xAI also came under fire for its environmental impact. The company’s new AI model, “Grok 3,” was trained at a “massive training center” in Memphis, raising significant concerns. The facility’s energy consumption was so high that it required three times the amount of electricity that the local utility could supply. On top of that, the center drew 30,000 gallons of water daily from the already strained Memphis Sand Aquifer. This raised concerns about overburdening the city’s already fragile infrastructure.
AI’s Impact on Water Resources:
Data centers require a lot of water for cooling the servers and other high-performance hardware to prevent them from overheating. For instance, Microsoft’s data centers in Goodyear, Arizona, are expected to use over 50 million gallons of drinking water annually, contributing to a water shortage in the region. Similarly, AI infrastructure in China consumes 1.3 billion cubic meters of water per year—almost twice the amount used by the city of Tianjin, which has 13.7 million people.
What about India?
India is also witnessing a data center growth, spurred by the rise of AI and digital consumption. There are now over 150 data centers in the country, run by major tech giants like Amazon, Google, and Microsoft. Most of these centers are concentrated in metropolitan areas like Mumbai, which alone accounts for more than 50% of India’s data center capacity.
To meet growing demand, India plans to double its power generation capacity to 820 GW by 2030. However, this rapid expansion of data centers raises critical concerns about energy consumption, water usage, and environmental sustainability, as the demand for electricity and cooling systems continues to strain resources.
Conclusion:
AI’s rapid expansion brings undeniable technological progress, but its environmental footprint cannot be ignored. The massive energy consumption, carbon emissions, and water usage associated with AI-driven data centers pose serious sustainability challenges. As demand for AI grows, tech giants must take urgent steps to mitigate these impacts—whether through renewable energy investments, improved efficiency measures, or stricter regulations on resource use. Without decisive action, AI’s promise of a smarter future may come at an unsustainable cost to the planet.