The Hidden Cost of AI: 477 Data Centres and Counting Across the UK

Across the UK — and particularly the south-east of England — giant, featureless buildings are springing up. No signage, no branding, deliberately unobtrusive. There are already 477 of them, with another 100 due in the next five years. They're not warehouses. As Countryfile presenter Tom Heap put it on BBC Radio 4's Rare Earth, they're "computers in sheds": millions of servers in racks, in huge barns, owned by the likes of Amazon, Google and Microsoft.

What's driving the boom?

Artificial intelligence. Modelling human language and behaviour demands an enormous amount of computing power — which means enormous amounts of electricity, land and water. Researchers at MIT estimate a single ChatGPT enquiry consumes about five times more electricity than a simple web search. Training a large language model like OpenAI's GPT-3 uses electricity equivalent to the annual consumption of 130 US homes.

"Computers are in essence just incredibly efficient electric heaters," says Heap. "For every ChatGPT query you make, it takes one pint of water to cool the system."

The numbers behind the hype

477 data centres already standing across the UK, with ~100 more planned in the next five years.
£10 billion, 54 hectares: a single new complex of 10 structures planned for the former Blyth Power Station site, with work due to start in 2031.
2–4% of the world's electricity is consumed by data centres — and the figure is rising fast as AI spreads.
175 billion parameters in ChatGPT-3.5; up to 1.8 trillion in ChatGPT-4 — training it produced emissions perhaps 12 times higher than its predecessor.
20 AI systems in 2022 emitted more carbon than 137 individual countries, according to researchers at Zhejiang University.

Why it matters locally

Every one of those UK data centres needs a grid connection, a water supply and a planning approval. Communities from Buckinghamshire to Blyth are discovering that the "computer in a shed" next door is a 24/7 industrial facility — with transformer hum, cooling towers, generator testing and construction traffic, often approved through the industrial-reuse loopholes we've covered before.

The environmental picture is no better: energy-greedy AI is a major and growing contributor to climate change. The Soil Association's Adrian Steele is blunt: "It is boosting the production and consumption of fossil fuels at a gigantic scale."

The counter-argument

The big tech companies are aware of the problem. Google uses AI algorithms to optimise data-centre cooling and cut energy use. Microsoft trialled underwater data centres off Orkney between 2018 and 2020, where the sea did the cooling. And more electricity is coming from renewables.

There's also genuine potential for AI to help — modelling weather and pests for farmers, improving soil health without chemical fertilisers, advancing climate modelling. As one Zhejiang University researcher put it: "AI has the potential to significantly benefit the environment... but it is a double-edged sword."

The bottom line

The AI boom isn't weightless. Every query you type has a physical cost — in electricity, in water, and in the quiet industrial buildings appearing in communities that never asked for them. As the UK's 477 sheds become 577, the question isn't whether AI is useful. It's who pays for what it consumes — and who gets to decide where the next building goes.

Source: Countryfile — The Cost of AI · Part of the Brightfull Tech network. Contact: hello@brightfull.tech