Discover AI’s real carbon footprint – and how the same tech is boosting solar, wind, and battery-storage efficiency.
Artificial intelligence (AI) is everywhere right now, and it seems we’re using it for everything, too. Recipes, funny cat pictures, bizarre videos, we just can’t get enough of the technology. But is there an environmental drawback to all this?
In short, yes, AI guzzles electricity like it's going out of style – though perhaps you haven’t heard about how it can be used to help the environment.
We’re giving you the definitive answer on whether AI is bad for the environment. We’ll demystify not only the negative impacts of AI on the environment, but also its massive potential for saving it.
PS We offer MCS-certified solar panel installation and battery installation nationwide. Simply answer these questions, get your fixed price, and arrange your free design.
🔑 Key Points:
Searching with AI uses more electricity than standard search engines.
Data centres, the backbone of AI, need millions of litres of water to stay cool.
AI technology can actually be hugely beneficial to the environment.
Solar panels and battery storage can both benefit from incorporating AI.
More efficient AI models are being developed that’ll use less energy.
What Is the Environmental Impact of AI?
The main environmental impacts to consider when talking about AI are the carbon emissions from electricity usage, water demands for data centres, and the materials needed for hardware.
In terms of electricity, AI, for all its capabilities, is heavily reliant on vast data centres to power its ability to generate what we ask it to. These data centres require a lot of energy to operate, which naturally means AI is consuming a chunk of electricity to function. Subsequently, every time we ask AI to do something or when a developer trains an AI model, we’re placing demands on the grid to make this possible.
Data centres need to remain active 24/7, so this demand for electricity never gets a break. Here’s a few mind-boggling stats about the amount of electricity AI uses:
Every time AI generates an image, it uses around 1,700 joules (enough to power a laptop for 30 seconds).
A ChatGPT query is estimated to use around five times more electricity than a normal internet search (3.6 joules vs 0.72).
There are around one billion queries processed on ChatGPT every day
In 2024, data centres electricity consumption around the world was estimated to be 415 terawatt hours (TWh). This could rise to 1,050 TWh by 2026.
By 2030, AI usage in the UK alone could reach 72 TWh.
The National Grid has issued a warning that AI use might increase the UK’s total electricity demand by 500% in the next decade.
For water, the data centres powering AI are incredibly thirsty, because they need water to keep cool. Your average data centre can use between 11 million and 19 million litres of water per day, which is roughly enough to meet the water needs of between 30,000 and 50,000 people.
With 523 data centres in the UK (as of 2025), that’s 7.84 billion litres of water keeping the key components cool.
This means data centres are using enough water to meet the needs of 55.2 million UK citizens every day (we’re assuming the data centres are using the average figure of 15 million litres of water).
Finally, the hardware used in the data centres often requires rare earth materials, such as yttrium, and cobalt, which is a critical component in the semiconductors fundamental to AI’s ability to process information and complex computations.
Unfortunately however, mining cobalt has devastating effects on the environment, including:
Contaminating water and crops
Killing wildlife
Polluting the air around the cobalt mines
Threatening the health of the humans that mine it
We’re wholly reliant on cobalt for more than just AI – smartphones, electric cards, laptops, and even the engines of aeroplanes all need cobalt to function. For all the ills that come with mining the material, cobalt isn’t going away any time soon.
The Carbon Footprint of AI
The intense electricity demands of AI means it’ll inevitably be responsible for some carbon emissions. Exactly how much is difficult to say, but we can draw a few conclusions from the available data.
Generating images with AI for example, uses approximately 1,700 joules, which translates to 0.0004722222 kWh. Around 34 million images are generated each day, requiring approximately 16,055 kWh.
In terms of emissions, the UK's grid electricity emission factor is around 0.22499 kg CO2e per kWh, so every day generating images with AI releases 3,612 kg of CO2e into the atmosphere. Annually, generating images with AI will release around 1,318,380 kg of CO2e.
Considering that the average CO2e per person in the UK varies between 5 tonnes and 12.7 tonnes of CO2e per year, it’s wild how much AI emits just on generating images alone.
If we use the estimated figure of AI using 72 TWh per year by 2030, AI will be responsible for emitting some 16,199,280 tonnes of CO2e annually, or 3.8% of the entirety of the UK’s carbon emissions.
A quick side note on the term “carbon footprint”; as important as it can be to track emissions, a “carbon footprint” was actually coined by an advertising firm working for BP. The idea was that it’d shift the onus of climate change and carbon emissions onto the individual, rather than it being the fault of oil giants.
Ways AI Can Be Used to Help the Environment
AI, despite the problems with high energy usage and emissions, does actually have huge potential for benefiting the environment, when it’s used in the right way. Experts are using the technology to pioneer innovative solutions to environmental challenges that’d otherwise be impossible or at least very difficult to solve without AI.
Let’s take a look at some of the ways AI is helping the environment:
It Can Track Air Pollution
Globally, air pollution contributes to some 8.43 million premature deaths every year. It’s so bad that the World Health Organisation ranks air pollution as the biggest environmental threat to humans worldwide.
So tracking it is absolutely essential to preventing people from breathing in air pollution unknowingly. Up until AI, this was done with monitoring stations needing extensive data points, which takes a lot of time.
Now, an AI algorithm developed by engineers at Cornell University can rapidly speed this process up. The model uses data from traffic, meteorology (the study of the atmosphere), and topology (a type of geometry) to run simulations that can help pinpoint when and where air pollution will spike.
AI Is Helping Clean the Ocean
The Ocean Cleanup, an environmental organisation from the Netherlands, is using AI as part of its mission to reduce plastic pollution in the world’s oceans.
By using AI to create maps of ocean litter, The Ocean Cleanup can more precisely target the most affected locations and remove the plastic waste. This process of using AI is much more efficient (and accurate) than the older method of using trawlers and aeroplanes.
Deforestation Can Be Tracked and Prevented
We lose roughly 10 million hectares of forest every year, which is nearly equivalent in size to the entirety of Iceland. Losing forests means losing irreplaceable ecosystems and exacerbating climate change by removing the forests’ ability to absorb carbon emissions.
Experts are coming up with creative uses for AI to help prevent this, with clever techniques such as combining AI with satellite imagery to map and track areas most affected by deforestation. This method can also help spot where illegal logging is happening and identify at-risk areas by using predictive models.
Using this predictive model to track land-use patterns, infrastructure development, and local populations, conservation specialists can better plan support for forests. Clever stuff!
It’s Decarbonising Industries
We’re still emitting billions of tonnes of CO2e each year, and industry is one of the worst offenders – the UK’s industry sector is estimated to have been responsible for 13% of all the country’s emissions in 2024.
Some 30% of all global emissions come from industry, so decarbonizing it to mitigate its impact on the environment is essential.
Thankfully, AI has some pretty creative solutions:
Improving supply chains. Supply chains are notoriously complex and difficult to optimise, leading to inefficiencies and higher carbon emissions. AI can simplify this, helping industries identify opportunities to make transportation routes more efficient, improve energy efficiency, and reduce waste.
Helping the shift to renewable energy sources. Industries are heavily reliant on fossil fuels, but AI can optimise energy systems and help industries incorporate renewable energy more effectively. Examples include identifying what parts of an industry can be powered by solar panels.
Making buildings more efficient. AI can be used to assess and monitor the energy systems of buildings, looking at energy usage, heating, cooling, and lighting. It can model patterns and help businesses make changes to improve efficiency, reduce wastage, and shrink energy bills.
AI Is Making Recycling Easier
The OECD (Organisation for Economic Co-operation and Development), estimates that only 9% of the world’s plastic waste gets recycled. Plastic pollution is one of the worst impacts of the modern world and we need to drastically improve how we manage it.
And, you guessed it, AI is making this easier.
Technology such as smart recycling bins and AI-powered recycling robots are improving the way we recycle waste. The bins, for example, tell people when they’ve put in an object that can’t be recycled, or needs to be recycled elsewhere.
Then there’s intelligent waste recovery powered by AI, which is being used to help recycling centres recycle more materials and reduce the amount going to landfill.
Climate Disasters Can Be Better Predicted
While we want to limit climate change as much as possible, it’s inevitable that extreme weather events will become more common, so figuring out a way to predict them will be more important than ever.
State-of-the-art AI modelling is doing exactly that, with examples such as GraphCast from Google being able to work out when devastating weather events such as cyclones might happen, with a lot more accuracy than traditional methods.
AI Makes Solar Panels Smarter
Solar farms are rapidly becoming one of the world’s biggest sources of renewable energy, and AI is helping make them more effective.
For example, just above we mentioned AI’s capability to predict weather patterns – well, that’s ideal for helping solar farms maximise output and in helping grids manage spikes in electricity generation.
AI can also be used for predictive maintenance. Measuring factors such as temperatures, irradiance, orientation, tilt angle, humidity, rainfall, dirt accumulation, power output, inverter efficiency, and operational loads, AI can make it easier to predict when solar panels require maintenance. This, in turn, reduces downtime for solar farms.
Read more:
What Are the Advancements in Solar Panel Technology? (2025)
AI Battery Storage is Much Smarter Too
Battery storage, especially on the larger scale, is utilising AI to store and release energy more efficiently. This is going to be particularly important as renewable energy continues to make up a bigger share of the National Grid.
We’ll need large-scale battery storage, because as good as renewable energy is, sources such as solar and wind power can still be intermittent. We need to be able to store it in giant batteries for use when either the sun isn’t shining or the wind isn’t blowing.
Read more:
Farming Is Becoming More Efficient With AI
Keeping the world fed and reducing agriculture’s impact on the environment is a massive challenge.
AI-powered farming is already rising to this challenge, with big steps taken to help farmers use resources more efficiently and have less of a negative effect on the land. Irrigation technology optimised by AI works to deliver the right quantities of water to crops, for example.
Algorithms are also capable of accurately analysing soil conditions and crop health, which help with optimal planting and harvesting respectively. And AI is being used in livestock management to check up on animal health and behaviour, improving welfare and reducing waste.
Something else that can help minimise agriculture’s environmental impact is using AI to predict food demand. By knowing and understanding the current demand for food, supply chains can be much more efficient and food waste will go down.
Can We Reduce the Impact of AI on the Environment?
If, as the International Energy Agency (IEA) predicts, AI reaches 4% of global annual energy usage by 2026 (more than the entirety of Japan’s energy usage), its impact on the environment will be difficult to ignore. Sure, the technology does have a lot of potential for helping the environment, but it’s important that AI’s damage doesn’t outstrip the benefits it could bring.
So how do we reduce the negative impact on the environment? For starters, making AI models more efficient is paramount. The more efficient an AI model is, the less energy it needs to achieve the same level of accuracy as a more energy-intensive, inefficient model.
Then there’s something called LLM Distillation, which is a technique aiming to transfer the learnings of a large pre-trained model to a smaller model. Training an AI model uses a lot of computational power and therefore a lot of energy – being able to reduce or even skip this step will shrink AI’s potential for damaging the environment.
There’s a few steps you can take too, including:
Try to avoid using AI when you don’t need to. AI can be useful for so many things, but take a second to ask yourself if you really need to use AI for whatever it is you’re about to do. Do you need to generate yet another image of a cat wearing full plate armour? Or should you ask AI to rewrite your casserole recipe as Yoda?
Think carefully about your prompts. If you need to use AI, take a moment to think about your prompt before entering it. Less is more after all, and reducing the number of prompts you use to get the desired result will mean using less energy.
Summary
We’re at a crossroads right now with AI. On one hand, there’s no denying its vast hunger for energy and water. In a world transitioning away from fossil fuels, do we need AI putting more strain on the grid? And should we divert precious drinking water to cool the growing number of data centres?
On the other, AI’s potential for helping with and solving critical environmental issues is colossal. Through responsible applications, AI could have a transformative effect in reducing or even reversing the worst impacts of climate change.
It’s a balancing act, and one that needs to be approached with care.
