One of the most heralded forces of progress—artificial intelligence (AI)—is subtly turning into a serious environmental threat in a world where climate change is a pressing concern.
Tech companies frequently ignore a developing crisis as they brag about advances in generative models, real-time language translation, and AI-driven healthcare: AI’s startling energy consumption and the carbon footprint of the data centers that support it.
According to a June 2025 report by the United Nations’ International Telecommunication Union (ITU), operational carbon emissions of the world’s leading tech companies rose by an average of 150% between 2020 and 2023.
The criminals? AI is developing quickly, and data center capacity is expanding exponentially.
For instance, compared to their 2020 statistics, Amazon’s carbon emissions increased by 182%, Microsoft’s by 155%, Meta’s by 145%, and Alphabet’s (Google’s parent company) by 138%. These figures are unacceptable in addition to being startling.
The Hidden Cost of the AI Assistant
We are using a huge, energy-intensive infrastructure each time we engage with an AI chatbot, generate customized content, use AI to illustrate a Ghibli style picture or organize a trip. Large server farms that run on non-renewable energy are the foundation of these services, necessitating constant cooling and processing.
Training a single AI model already has a huge environmental cost. According to a 2023 MIT study, over the course of their lifetimes, training a large language model like GPT4 emits as much carbon as five average cars. The cumulative effect of training and retraining hundreds of models annually is astounding.
However, tech firms frequently hide these facts behind selective reporting and ambiguous sustainability pledges. According to the ITU report, “there are currently no international standards or legislative requirements for companies to disclose their AI-specific energy usage or emissions.” We are in the dark if there is no transparency.
A Boom Fueled by Coal and Gas
The irony is that many tech leaders brand themselves as climate-conscious innovators. Google claims to run all its operations on renewable energy, while Microsoft has pledged to be carbon negative by 2030. However, these commitments often exclude indirect emissions—like those from the third party data centers or electricity providers still reliant on fossil fuels.
According to the International Energy Agency (IEA), data centers consumed 415 terawatt hours (TWh) in 2024, accounting for about 1.5% of global electricity demand—more than some entire nations. And this figure is projected to more than double by 2030, possibly exceeding the total electricity consumption of Japan.
Worse, the demand is concentrated among a few powerful firms. Just 10 companies—including Amazon, Microsoft, Meta, and China Mobile—accounted for over 50% of the total electricity consumption among the top 200 digital firms in 2023, according to Statista.
This concentration of power also means these companies have outsized control over global emissions, yet are subject to virtually no global environmental governance.
Tech Climate Accountability Gap
One of the most alarming insights in the ITU’s report is not just the rise in emissions, but the lack of global oversight. Despite international accords like the Paris Agreement, there is still no binding mechanism that holds digital corporations accountable for their climate impact.
Regulators have been sluggish to adjust. Governments discuss how to control the ethical risks of AI, such as bias or false information, but they rarely consider the environmental risks. Rarely do the same governments that fund digital infrastructure demand energy caps or eco-audits for these platforms.
Tech companies frequently release self-congratulatory sustainability reports that are chock-full of selective data and aspirational goals. The distinctions between Scope 1 (direct), Scope 2 (indirect), and Scope 3 (supply chain-related) emissions are rarely made explicit in these documents. The outcome? a positive image that frequently ignores the actual environmental cost of innovation.
Is Green AI a Thing?
Hope exists, but only if quick action is taken. Low-energy AI models are being developed by some researchers, and there is increasing interest in moving away from large, distant servers and toward edge computing, where data processing takes place closer to the source (such as on your phone).
As part of their climate strategies, countries such as Germany are starting to examine data center energy consumption. But this cannot be resolved by innovation alone. We require enforcement, disclosure requirements, and regulation. Digital infrastructure must be held to the same standards as sectors like manufacturing and aviation, which are subject to carbon taxes and emissions caps.
The complete transparency of how these businesses power their platforms must be demanded by investors, consumers, and civil society. Imagine product labels or app store ratings that include not just user reviews but “carbon scores”—a simple way to let users know the environmental impact of using a particular AI tool.
Innovate Consciously or Perish
Without a doubt, one of the most revolutionary technologies of our day is artificial intelligence. It can enhance medical diagnostics, transform education, and aid in the prediction of natural disasters.
However, advancement must not come at the planet’s expense. We must choose between creating a future where sustainability and artificial intelligence coexist or letting the digital age subtly hasten the demise of the environment.
We must consider the impact AI is having on the planet as we are in awe of its capabilities. We are paying an excessive price if the price of intelligence is a burning planet.
Arshad Ali
Arshad Ali is a student of International Relations and an English language instructor. With an interest in international affairs, diplomacy, and literature, he positions himself as a dynamic voice in both education and global discourse.
