Voix du Paysan: educating and informing citizens — a commitment to social and climate justice.
Behind the promises of artificial intelligence lies a less virtual reality: data centres that consume vast amounts of electricity, water and land. The United Kingdom wants to triple its capacity by 2030. But at what environmental cost?
The United Kingdom dreams of becoming an “AI superpower”. To achieve this, London plans to increase the capacity of its artificial intelligence data centres to 6 gigawatts, up from around 1.8 GW today. Behind the screens and algorithms, however, lie colossal infrastructures: warehouses packed with servers that consume enormous amounts of electricity and require substantial quantities of water for cooling. Is AI therefore becoming the new face of a form of progress that claims to dematerialise the world while simultaneously increasing its physical footprint?
The paradox is stark. At a time when the United Kingdom is seeking to cut emissions and adapt to droughts, some data centres can consume millions of litres of water every day. At the same time, hundreds of projects are awaiting connection to the electricity grid, with a combined capacity far exceeding the country’s average demand. AI does not float in a cloud: it relies on machines, cables, power stations, cooling systems and energy infrastructure whose environmental cost cannot be erased by a few lines of code.
The technology industry argues that AI could become a weapon in the fight against climate change: optimising electricity grids, reducing waste, improving weather forecasting and accelerating energy innovation. Why not? But a promise is not proof. When the climate benefits being touted rely more on commercial projections than on robust scientific assessments, caution becomes an obligation. Especially when, at the same time, AI’s growing energy demand is driving some markets to extend or build fossil-fuel generation capacity. A technology does not become “green” simply because its rhetoric is.
The real debate, therefore, is not whether we should be for or against artificial intelligence. It is about what kind of AI we want — and who should bear its environmental cost. If data centres monopolise electricity, appropriate water resources and undermine climate targets, economic growth cannot be used as a licence to pollute. The United Kingdom must impose strict rules: transparency over water and energy consumption, comprehensive carbon assessments, priority for genuinely decarbonised energy sources, and protection of local resources. Innovation cannot be an excuse for avoiding restraint.
For behind every request made to a machine now lies a political question: how many resources are we willing to sacrifice so that artificial intelligence can appear limitless? The danger would be to repeat with AI the mistakes made with fossil fuels: pursuing power at all costs, only to discover the environmental bill too late. True progress will not be measured by who builds the most data centres, but by who succeeds in aligning digital intelligence with planetary boundaries. An AI capable of calculating the world should also be capable of understanding that it depends on it.
Emmanuel Ndimwiza
Editor — Voix du Paysan
AI: When Digital Progress Dries Up the Planet
