It is difficult to obtain reliable figures on the climate impact of AI

A key challenge is that tech companies rarely disclose how much energy their AI services actually consume. This makes it difficult—for businesses, researchers, and policymakers—to assess the climate impact of a single AI query. What we do know is largely based on estimates and modeling from independent research.

This is how much electricity AI and data centers consume

Behind every AI query, every internet search, and every streamed movie lies a physical structure: a data center. These are large facilities filled with servers that store and process digital information around the clock.

Data centers consumed approximately 415 terawatt-hours (TWh) of electricity globally in 2024—equivalent to about 1.5 percent of the world’s total electricity consumption. AI is the single most important driver of this growth. According to the IEA, electricity consumption by data centers is projected to more than double to around 945 TWh by 2030 (IEA, 2025).

To put this into perspective: Google has compared the energy consumption of sending a text query to an AI to watching TV for 9 seconds (Elsworth et al., 2025).

AI servers in the U.S. could account for 10 percent of the country's electricity consumption by 2030

A study published in *Nature Sustainability*, involving researchers from institutions including KTH, Cornell, and Oxford, estimates that by 2030, AI servers in the United States alone could consume up to 10 percent of the country’s current electricity usage—an amount comparable to Germany’s total annual consumption (Xiao, Nerini et al., 2025).

The study also shows that the energy mix plays a major role: a faster transition to renewable electricity could reduce emissions from the AI sector by more than 15 percent, while a slower pace risks increasing them by about 20 percent.

AI's carbon footprint and water usage

The climate impact of AI is not limited to electricity. A study by de Vries-Gao (2026) estimates that the total carbon footprint of AI systems in 2025 ranged from 32.6 to 79.7 million metric tons of CO₂e. In addition to greenhouse gases, research points to a significant water footprint: the total water consumption of AI systems is estimated at 312.5–764.6 billion liters in 2025—comparable to the global annual consumption of bottled water.

The need for greater transparency regarding the environmental impact of AI

A key methodological challenge is that neither the IEA’s statistics nor the sustainability reports of tech companies currently distinguish AI-specific workloads from other data center operations. This makes monitoring and comparability difficult. For climate action to keep pace with AI development, clearer accounting standards and better data are needed.

ZeroMission developments on an ongoing basis.

Do you factor in the use of AI in your climate reporting?

This is an issue many sustainability managers are grappling with right now. Contact one of our climate strategists—we’ll help you manage Scope 3 emissions, data collection, and reporting in accordance with CSRD the GHG Protocol.

Sources: Elsworth et al. (2025), “Measuring the Environmental Impact of Delivering AI at Google Scale,” arXiv:2508.15734; IEA (2025), “Energy and AI,” World Energy Outlook Special Report; KTH Climate Action Centre (2025), on Xiao, Nerini et al. (2025), Environmental impact and net-zero pathways for sustainable artificial intelligence servers in the USA, Nature Sustainability de Vries-Gao, A. (2026), The carbon and water footprints of data centers and what this could mean for artificial intelligence, Patterns 7, CellPress Luccioni, S., Jernite, Y. & Strubell, E. (2024), Power Hungry Processing: Watts Driving the Cost of AI Deployment?, ACM FAccT 2024