Global race to build AI data centres raises energy, resource and policy challenges

Governments and tech firms worldwide are competing to erect massive AI data centres, a trend that is reshaping energy use and prompting policy debates. Former US President Donald Trump dismissed local opposition to data‑centre construction as foreign‑influenced politics, urging communities to accept the projects. In the United States, BloombergNEF predicts data‑centre electricity demand could reach 20 % of national generation by 2035, with AI training workloads accounting for roughly half of that load. India’s Ministry of Power forecasts electricity needs of up to 13.6 GW for data centres by 2031‑32, while a Mumbai conference highlighted the need for sustainable power, cooling and skilled labour. Argentina’s “Stargate Argentina” proposal aims to leverage Patagonia’s renewable energy and cool climate for a mega‑centre, though only a letter of intent with OpenAI has been signed. Australia’s leaders argue that future AI infrastructure should shift northward to exploit abundant renewable resources and reduce pressure on the southeast’s water and power networks. Thailand is piloting a direct power‑purchase agreement that lets large data centres contract renewable generation instead of natural gas. Across the globe, community protests, water‑use concerns and calls for national siting strategies underscore the growing strategic importance of AI data‑centre development.