(CNN) — The rapid proliferation of artificial intelligence (AI) has sparked significant political debate, centering around the environmental impacts of the necessary data centers. These vast structures, essential for AI’s operation, have grown exponentially, raising alarms about their energy and water consumption, as well as the noise and heat they generate.
A CNN poll conducted by SSRS in September reveals that nearly 90% of Americans express more concern than excitement over the rise of AI data centers. The environmental implications of these facilities are viewed by some as catastrophic, while others dismiss these fears as exaggerated or even unfounded.
The environmental impact of AI is complex, influenced by factors such as location, energy source, and technology used in data centers. A lack of comprehensive and transparent data from AI providers complicates the situation further, leaving us reliant on estimates and projections.
High Energy Consumption
AI requires significant computing power, leading to substantial energy demands. Despite improvements in AI’s energy efficiency, the surge in usage and energy-intensive applications like video generation has led to a 50% increase in electricity consumption by AI data centers last year, according to the International Energy Agency (IEA).
In 2025, AI data centers accounted for 0.5% of global electricity consumption, a figure projected to triple by 2030, as per an analysis from Our World in Data. The United States alone could see data centers consume up to 12% of its electricity by 2028, partly due to the rapid expansion of AI, based on an analysis from Lawrence Berkeley National Laboratory.
Climate and Health Concerns
With electricity generation still heavily reliant on fossil fuels, the environmental toll of AI data centers includes significant contributions to climate change. While some power is sourced from renewables, coal remains a primary source at 30%, followed closely by renewables and natural gas, as outlined in a report by the IEA.
Natural gas is a growing source of power for data centers in the US, with a substantial portion of new gas-fired power capacity intended for AI, according to a report from Global Energy Monitor.
The public health impact in the US could reach $20 billion by 2028, due to emissions from gas turbines used by data centers, as reported in a recent study.
Water Usage
AI data centers are also significant water consumers, both directly for cooling and indirectly through electricity generation. Estimates vary, with a single AI-generated email potentially using as much as a bottle of water, or much less, according to OpenAI CEO Sam Altman, who quotes about one fifteenth of a teaspoon.
A March analysis by scientists from the University of California suggests US data centers could use 110 billion gallons of water annually by 2030. Additionally, a report from Ceres highlights that data centers across seven states currently require about 3.4 trillion gallons of water annually for electricity production.
Noise and Heat Pollution
Data centers can disrupt local communities with noise pollution from cooling systems and power generators. Residents, such as those near a large data center in Vineland, New Jersey, have filed lawsuits over noise concerns.
Furthermore, these facilities might contribute to urban heat islands, as a study from Arizona State University found, with data center heat potentially raising local temperatures by several degrees.
Mineral Demands and E-Waste
The growing AI industry requires significant amounts of critical minerals for semiconductors and data storage, often mined in the Global South, leading to environmental and health issues. With AI’s growth, electronic waste is also expected to increase, yet only about 20% is currently recycled.
AI’s mineral demands may compete with renewable energy sectors, which are vital for addressing climate change.
The Role of Human Decisions
Ultimately, the environmental impact of AI hinges on human choices, including the placement of data centers, their power sources, and regulatory measures. According to Kaveh Madani of the United Nations University Institute for Water, Environment and Health, it’s crucial to assess the benefits and trade-offs before constructing new AI data centers.
Madani emphasizes, “There are trade-offs like any other technology. It has downsides that need to be regulated, monitored, and mitigated.”
Original Story at www.wxow.com