AI data centers to consume 3X combined annual electricity use of Pakistan, Bangladesh and Nigeria by 2030: UN

A recent UN study warned of side effects of AI boom, projecting that data center power use will double and water use will rise by 9.3 trillion liters by 2030.Is AI boom eating away natural resources? UN report warns of carbon, water and land footprint of running data centers.(REUTERS)On the occasion of World Environment Day, United Nations University published a report suggesting the environmental impact of running data centers. From water and land footprint to power consumption, these global data centers powering artificial intelligence are projected to consume nearly triple the combined annual electricity use of Pakistan, Bangladesh and Nigeria by 2030. To meet the surge in demand from artificial intelligence, it is expected that the water footprint will equal the basic annual domestic water needs of all 1.3 billion people in Sub-Saharan Africa.The research published on June 3 by UN Institute for Water, Environment and Health warned against the rising ​environmental costs of AI as the rapid rollout could ⁠also strain scarce land resources and create mountains of electronic waste. The researchers emphasized that environmental costs of AI and data centers should not be considered through carbon emissions in isolation as it has far-reaching impact. Quantifying carbon, water and land footprints of AI's electricity use across the globe, the study underscores the major differences between these footprints in the world’s 20 largest data center hubs.Investigation team leader and Director of UNU-INWEH Professor Kaveh Madani said, “This report is not a case against artificial intelligence, a technological transformation that is improving the lives of billions of people around the world. It is a call for using it responsibly and addressing its unintended impacts proactively to make it sustainable and equitable."Raising calls for sustainable and responsible use of natural resources for advancing AI, he added, "We have a narrow window to ensure that the backbone of the technological revolution of our era develops within planetary limits, and that the communities who provide the critical minerals for advancing AI and the ones that host its infrastructure and e-waste are also among those who benefit from it."Describing why environmental costs of AI and data centers should take into account water footprint and land footprint in addition to carbon emissions, the report observed that most existing assessments are systematically mismeasured. The researchers found that AI sustainability measured through a single metric can hide trade-offs and shift environmental burdens. Notably, the above mentioned three footprints do not move in the same direction as "low-carbon" emission corresponds a higher water and land footprint.Elaborating on the key findings UNU-INWEH Researcher and the lead author of the report, Dr Miriam Aczel, said, “What surprised us most is how often the choices that look greenest from a carbon perspective end up worse for water or for land.” She added, “If we keep judging AI sustainability by carbon alone, we might think that renewables make AI infrastructure clean but that is solving one problem while creating other problems, often in places that didn't ask for it.”Key findings of UN report:Data centers consumed 448 terawatt-hours of electricity globally in the previous year, ‌more than the whole of Saudi Arabia while AI accounted ‌for ⁠a fifth of the total.Last year, data centers consumed 4.5 trillion liters of water and generated 189 million ​tons of carbon dioxide emissions. Notably, data centers' water usage is enough to meet the needs of more than 600 million people in Sub-Saharan Africa.“The ‌public debate still often treats AI as software, but AI is also physical infrastructure: data centres, electricity generation, cooling systems, transmission networks, chips, minerals, land and water,” Kaveh ‌Madani noted.The report projects that the annual power consumption of data centers will double to 945 TWh by 2030. This figure i