AI and the environment
Our position. Last updated: August 2026
AI has an environmental cost and we don't pretend otherwise.
We also don't think the answer is for small businesses in Thanet to opt out while everyone else gets on with it. Most of that cost sits with the companies building the infrastructure, not with a café owner using AI to manage their suppliers. But how you use these tools does make a difference, and part of our job is teaching people to use them deliberately rather than wastefully.
The cost is real
Data centres (the warehouses full of computers that AI runs on) use electricity and water. A lot of both, and more every year.
The International Energy Agency expects global data centre electricity use to roughly double by 2030. In the UK, data centres already account for around 2.5% of all electricity, and the National Energy System Operator expects that to rise four-fold by the end of the decade.
The International Energy Agency puts global data centre water consumption at around 560 billion litres in 2025, roughly two-thirds of it from electricity generation.
What isn't settled
Plenty. Anyone who tells you they know exactly what AI will cost the planet is guessing.
- There is no comprehensive global dataset on data centre energy or water use. Almost no government requires operators to report it. Every figure you read is an estimate, and credible estimates vary widely.
- Efficiency is improving quickly. The energy used per AI task is falling. Whether that improvement outpaces the growth in demand is an open question.
- UK data centres generally use less water than American ones, because our climate allows less water-intensive cooling. Industry says that makes the water concern overstated. Campaigners say the real problem is that operators aren't required to disclose usage, so nobody can check. Both of those points can be true at once.
Water, and why it matters here
Kent is one of the driest parts of the UK, and Thanet is specifically identified as water-stressed. We have had drought status, hosepipe bans and supply problems in recent years. That pressure is not going away.
Most UK data centres sit in the south and east of England. It is also worth knowing that, on Parliament's own assessment, water companies' long-term supply plans do not currently account for the water needs of data centres at all.
To be clear about the scale, though: nobody is proposing a hyperscale data centre in Thanet. The link between your use of a chatbot and the level in a Kent reservoir is real, but indirect, running through national grids and global infrastructure.
What it means for your business
The honest answer is that your individual usage, especially in a small or solo business, is a rounding error. The environmental question about AI is a question about infrastructure investment, planning policy and how we generate electricity. It isn't settled by whether you write your invoices by hand or with ChatGPT.
That said, waste is waste, and a lot of AI use is wasteful. The wrong tool for the job, the same request run twenty times, output nobody reads. Wasteful use is usually bad work as well. The habits that lower the environmental impact are the same habits that get you a better result, faster. That's what we focus on in all our training.
What we commit to
Right-sized tools
We teach people to pick the smallest, cheapest model that does the job, and to get a good result in two attempts rather than twenty.
Local delivery
We work in Thanet and East Kent. Sessions happen locally: at your premises, a local co-working space, or our studio in Margate.
Support for local good causes
Charities, community groups and schools in Thanet get reduced or free sessions. If you're doing good work locally and cost is the thing stopping you, ask us.
No greenwashing
We won't repeat AI companies' sustainability claims as though they were verified facts. Where the evidence is contested, we'll say so. Where we don't know, we'll say that too.
Read the evidence yourself
Don't take our word for it. These are independent, non-commercial sources. None of them is selling you AI.
- What are data centres and how sustainable are they?Parliamentary Office of Science and Technology. The clearest short summary of energy, water, e-waste and regulation in the UK.
- Data centres: planning policy, sustainability, and resilienceHouse of Commons Library. Sets out the UK water and energy debate, including where industry and campaigners disagree.
- Energy and AIInternational Energy Agency. The most thorough global modelling of AI's electricity demand, and the source most others cite.
- AI has an environmental problemUN Environment Programme. Covers water, minerals and e-waste alongside energy, and what governments could do about it.
- Five charts that put data centre energy use into contextCarbon Brief. Useful for scale: how AI's footprint compares with aviation, road transport and industry.
- Future water resourcesHouse of Commons Library. Background on the water supply deficit facing the south east of England.
Tell us if we've got this wrong
The evidence on this is moving quickly and we'll update this page as it does. If you think we've misread something, or you've found better sources, we'd like to hear it. Email Contact@Margate.Ai.