Generative AI is energy -intensive and the ways in which its environmental impact can be calculated are complex. Consider the downstream effect of generative AI on the environment as you examine your company’s own sustainability goals.
- What side effects may not be immediately visible, but can have a major influence?
- When does most of the energy consumption occur: during exercise or daily use?
- Does “more effective” AI models actually address any sustainability problems?
The effect of generative AI on electricity generation, water and air quality
AIS influence on air pollution
In December 2024, the University of California, Riverside and California Institute of Technology calculated that training Metas Llama-3.1 produced the same amount of air pollution as more than 10,000 round trips by car between Los Angeles and New York City.
The increased air pollution from backup generators on data centers running AI caused regional costs of public health of about $ 190 million to $ 260 million a year, researchers found in UC Riverside and Caltech.
AI’s influence on electricity consumption
A 2024 report from the International Energy Agency said a chatgpt prompt used 10 terawatt-hours more electricity a year than the total used annually for Google searches.
AI’s influence on water consumption
To crooked more electricity could already struggle for tools, leading to brownouts or blackouts. Drawing water from already drought exposed areas, such as the rapidly developing Phoenix, Arizona or the deserts of California, can cause loss of habitats and fires.
See: To send an E -Mail with Chatgpt equivalent to ingesting a bottle of water
Consumer training or daily use of AI more resources?
“Education is a time -consuming and energy -intensive process,” IEA wrote in its 2025 Energy and AI World Energy Outlook Special Report. A GPU of the type suitable for AI training draws about as much electricity as a toaster at the maximum assessed power consumption. The agency calculated that it took 42.4 gigawatt-hours to train Openais GPT-4, corresponding to the daily household electricity use of 28,500 households in an advanced economy.
What about everyday use? Inquiry size, model size, the degree of inference time scaling and several factors in how much electricity an AI model uses in the inference stage in use, to analyze the prompt. These factors and lack of data regarding the size and implementation of consumer -AI models mean that the environmental impact is very difficult to measure. However, generative AI draws undeniably more power than conventional computing.
“The inference phase (also the operational phase) was already responsible for the majority (60%) of AI energy costs on Google, even before mass recording of generative AI applications happened (2019-2021),” wrote Alex de Vries, founder of the DigiConomist and Bitcoin Energy Contbrug. “Although we do not have accurate numbers, mass uptake of AI applications will have increased the weight of the inference (/operational) phase further.”
Meanwhile, AI models continue to expand. “Increase the model size (parameters) will result in better performance, but increases energy consumption of both training and inference,” De Vries said.
Download: This GreenTech Quick Glossary of TechPublic Premium
Deepseek claimed to be more energy efficient but it’s complicated
Deepseeks AI models have been hailed to achieve as much as their biggest competitors without consuming so much energy and at a lower price tag; However, reality is more complicated.
Deepseeks Mixing of experts approach reduces the cost of treating relationships between concepts in batches. It does not require as much calculation power or consume so much energy during exercise. The IEA found that the daily use of the inference time scaling method used by Deepseek-R1 uses a significant amount of electricity. Generally, large inference models consume the most electricity. Education is less demanding, but the use is more demanding, according to my Technology Review.
“Deepseek-R1 and Openais O1 model are significantly more energy-intensive than other major language models,” IEA wrote in 2025 Energy and the AI report.
The IEA also pointed out the “rebound effect”, where the increased efficiency of the product leads to more users adopting it; As a result, the product continues to consume more resources.
Can AI equalize the resources it uses?
Technical companies still like to present themselves as good managers. Google is pursuing energy -conscious certifications globally, including the signature of climate -neutral data center pact in Europe. Microsoft, who saw similar increases in water and electricity use in its 2024 -smearitability reporting, is considering reopening a nuclear power plant at Three Mile Island in Pennsylvania to operate its AI data centers.
See: The proliferation of AI has created a lasting boom in data centers and related infrastructure.
Supporters of AI can argue that its benefits outweigh the risk. Generative AI can be used in sustainability projects. AI can help fight through massive data sets with information on carbon emissions or greenhouse gas emissions. In addition, AI companies are constantly working to improve the effectiveness of their models. But what “efficiency” really means that always appear to be the catch.
“There are some bottlenecks (such as grid capacity) that can hold back growth in AI and its power needs,” said De Vries. “This is difficult to predict, even considering that it is not possible to predict the future demand for AI (for example, the AI hype may fade to some extent), but any hope of limiting AI flow needs comes from this. Because of the” bigger is better “is dynamic AI fundamentally incompatible with environmental sustainability.”
Then there is the question of how far down in the impact chain AI’s impact should be counted. “Indirect emissions from the consumption of electricity are the most significant component of emissions from hardware making [of semiconductors,” said the IEA in the Energy and AI report.
The cost of hardware and its use has gone down as companies understand the needs of generative AI better and pivot to products focused on it.
“At the hardware level, costs have declined by 30% annually, while energy efficiency has improved by 40% each year,” according to Stanford University’s 2025 AI Index Report.
DOWNLOAD: This IT Data Center Green Energy Policy from TechRepublic Premium
Consider how generative AI affects your business’ environmental targets
Generative AI is becoming mainstream. Microsoft’s Copilot is included by default in some PCs; smartphone makers are eagerly adding video editing AI and assistants; and Google gives out its Gemini Advanced model for free to students.
Tech companies that set promising sustainability targets may find it difficult to hit their goals now that they produce and use generative AI products.
“AI can have dramatic impacts on ESG reports and also the ability of the companies concerned to reach their own climate goals,” said de Vries.
DOWNLOAD: This Customizable Environmental Policy from TechRepublic Premium
According to Google’s 2024 Environmental Report, the tech giant’s data centers consumed 17% more water than in 2023. Google attributed this to “the expansion of AI products and services” and noted “similar growth in electricity use.” Google’s data center waste generation and water use both increased.
“As AI adoption accelerates, IT leaders are increasingly aware that smarter devices don’t directly correlate to more efficient power consumption,” said Dan Root, head of global strategic alliances at ClickShare. “The spike in compute demand from AI tools means IT departments must look for offset opportunities elsewhere in their stack.”
As the International Energy Agency pointed out in its 2024 electricity report, both the source of electricity and the infrastructure need to be considered if the world is to meet the energy demands of AI.
“You can make/keep models a bit smaller to reduce their energy requirement, but this also means you have to be prepared to sacrifice performance,” said de Vries.