AI data centres are facing resistance across the world. In the US, opposition has evolved into a national movement. Last week, protesters gathered at more than 125 locations in the country’s first nationwide campaign against the rapid expansion of these facilities. In New York, lawmakers have proposed a one-year moratorium on new data centre construction while authorities assess the impact on electricity demand and the environment. Similar concerns are surfacing in Europe and Latin America as governments struggle to balance digital infrastructure with sustainability goals.
In Ireland, last year, the authorities ended a four-year moratorium on new data centre grid connections only after introducing strict conditions requiring facilities to generate power on-site and source most of their electricity from new renewable projects. The Netherlands has restricted hyperscale data centres to just two designated locations. In Chile, large facilities planned near Santiago have triggered community protests and legal challenges over concerns about water consumption and biodiversity.
India, too, is witnessing similar tensions. In Thane district near Mumbai, around 1,000 residents recently protested against Amazon’s proposed 422-megawatt data centre, arguing that such projects should be located in dedicated industrial or technology zones rather than close to residential neighbourhoods and schools.
Invisible Powerhouse
For those who came in late — every time you ask Google Gemini to draft an email, identify an image or answer a question, the lifting happens far from your phone or laptop. Your request is broken into tiny packets of data that travel through mobile towers, fibre-optic cables and the internet to data centres filled with thousands of high-performance computers. These facilities process the request, generate a response and send it back within seconds. While your device serves as the interface, data centres are the engines powering the AI revolution.
For decades, these facilities remained largely invisible. They powered cloud computing, streaming platforms, online shopping and enterprise software without attracting much public attention. But the rapid rise of generative AI has changed that. As technology companies invest hundreds of billions of dollars in AI infrastructure, data centres have become symbols of a growing debate over the environmental and social costs of AI.
The scale of the industry’s expansion explains the growing unease. The US has more than 6,000 operational data centres, accounting for over 40% of global capacity. Worldwide, there are more than 12,000 operational facilities, and the number is expected to rise sharply as AI adoption accelerates. India is also entering a major expansion phase. According to Rubix Data Sciences, the country’s installed data centre capacity is expected to grow more than fourfold by 2030. As of January, India had 271 operational data centres, with Mumbai, Hyderabad, Delhi-NCR, Bengaluru and Chennai accounting for nearly 65% of the total number. Global technology companies and domestic operators are investing heavily to meet rising demand for AI services, cloud computing and digital storage.
Power Grids, Scarce Water
Among the major concerns is the enormous amount of electricity AI requires. Unlike conventional internet services, generative AI models demand computing power for both training and responding to user queries.
Technology giants including Microsoft, Amazon, Google and Meta are investing tens of billions of dollars annually to expand their AI infrastructure. Consultancy McKinsey estimates that global demand for data centre capacity could more than triple by 2030, driven largely by AI. Modern AI data centres require as much power as a medium-sized city, while next-generation campuses are expected to exceed one gigawatt of electricity demand. In the US, Bloomberg estimates that data centres could account for around 20% of national electricity consumption by 2035, compared with just 5.9% today.
This surge is placing pressure on electricity grids and raising concerns that households and businesses may ultimately bear higher costs. Oracle reportedly faces more than $7 billion in collateral requirements for a planned Wisconsin data centre after regulators tightened rules designed to shield consumers from rising electricity expenses.
Water use has emerged as another critical concern. Thousands of powerful computer servers generate enormous amounts of heat and require continuous cooling. Many facilities rely on evaporative cooling systems that consume millions of gallons of freshwater, particularly during hot weather. In drought-prone regions such as Arizona, Nevada and Texas, residents argue that scarce water resources are being diverted towards AI infrastructure instead of homes or agriculture.
Communities are concerned about the broader impact on local infrastructure. Large data centres require new transmission lines, substations and additional electricity generation capacity, investments that can take years to complete. Environmental groups point to the conversion of farmland, forests and open spaces into sprawling industrial campuses, while residents complain about heavy truck traffic, construction noise and diesel generators that can affect local air quality during testing or power outages.
The economic benefits are also being questioned. Although construction creates thousands of temporary jobs, modern data centres are highly automated and employ relatively few permanent workers compared with factories occupying similar amounts of land.