The Environmental Costs of the AI Boom: A Closer Look at Emissions and Water Usage
The rapid rise of artificial intelligence (AI) has ushered in remarkable advancements across industries. However, it has also raised significant environmental concerns. A recent study by Dutch academic Alex de Vries-Gao reveals that the carbon dioxide emissions caused by AI in 2025 could rival those of New York City. This startling statistic underscores the hidden environmental costs associated with our increasingly digital lives.
Unprecedented Carbon Footprint
The findings indicate that AI-related activities are expected to generate a staggering 80 million tonnes of carbon emissions by 2025. To contextualize this, it equates to more than 8% of global aviation emissions. Such figures highlight the urgency of scrutinizing AI’s environmental impact, especially as the technology continues to permeate daily life through platforms like OpenAI’s ChatGPT and Google’s Gemini.
De Vries-Gao emphasizes that this is the first in-depth analysis targeting AI’s specific impact rather than the broader scope of data centers. As organizations embrace AI, understanding these environmental implications becomes crucial for fostering sustainability.
Water Usage: A Growing Concern
In addition to significant carbon emissions, the study estimates that AI operations will consume approximately 765 billion liters of water by 2025. This figure is particularly shocking as it surpasses global bottled-water demand. Previous estimates regarding water usage in data centers did not account for AI’s specific contributions. This finding marks a pivotal moment for environmental awareness, pushing the industry to reevaluate its water consumption strategies.
The consequences of this water usage are far-reaching. As AI technologies grow, the demand for cooling systems and other water-dependent infrastructures will only increase, leading to potential strains on essential water resources.
More Read
The Call for Transparency
The study urges technology companies to be more transparent about their environmental impacts. De Vries-Gao argues that society currently bears the agricultural costs of these advanced technologies while tech giants reap the benefits. With an estimated rise in datacenter electricity consumption expected to double by 2030, accountability becomes essential. Stricter regulations are necessary to ensure that companies are not only focused on innovation but also on the ecological footprint they leave behind.
De Vries-Gao’s research was published in the journal Patterns, highlighting that the International Energy Agency (IEA) estimates AI-focused data centers draw energy equivalent to that consumed by power-intensive aluminum smelters. As new technologies emerge, understanding their energy requirements must become a priority for developers and stakeholders alike.
Implications for Future Infrastructure
Donald Campbell from Foxglove highlights the potential for unprecedented emissions from future datacenters, particularly "hyperscale" facilities. Just one of these massive complexes can emit climate-changing gases akin to those produced by several international airports. In the UK alone, hundreds of these facilities are in the planning stages, leading to mounting concerns about their environmental repercussions.
The largest datacenter under consideration in the UK, planned at a former coal power station site in Blyth, Northumberland, is projected to emit over 180,000 tonnes of CO2 annually, comparable to the output of more than 24,000 homes. Such statistics paint a concerning image of what lies ahead if growth continues unchecked.
Global Energy Consumption Trends
As the demand for AI technologies rises, the most significant share of datacenter electricity consumption will continue to come from the United States, with China and Europe trailing. The trend raises questions about the sustainability of energy sources that power these data centers. Many regions are increasingly reliant on fossil fuels to meet demand, which exacerbates the overall carbon footprint of AI technologies.
In India, significant investments are flowing into datacenter construction. However, questions arise regarding reliability and the potential need for diesel generator farms as a backup power source, which consultants like KPMG warn could contribute to increased carbon emissions.
Navigating Corporate Responsibility
Technology companies are urged to improve their environmental disclosures to provide a clearer picture of their carbon footprint. De Vries-Gao points out that major players, such as Google, often fail to account for the total environmental costs associated with their AI systems. For example, while Google managed to reduce energy emissions from its datacenters by 12% in 2024, they acknowledge that the complexity of achieving climate goals has increased.
The growing disconnect between technological advancement and environmental responsibility requires urgent attention from all stakeholders involved.
Conclusion
The rapid expansion of AI technologies brings substantial environmental challenges that must be confronted. As we navigate this technological boom, highlighting the implications on carbon emissions and water usage becomes essential. The collaboration between tech companies, legislators, and the general public is crucial for fostering sustainable practices that balance innovation with ecological preservation.
Inspired by: Source
