Global datacentres are on track to produce 70% more carbon emissions by 2030, driven largely by the energy demands of artificial intelligence in the US and China, according to the International Energy Agency (IEA). The agency warned that datacentres will consume twice as much electricity in four years compared with current levels, with AI expansion accounting for three-quarters of that increase.
The IEA’s projection comes as a new draft text is set to be published during the ongoing conference this week, where the mood remains muted despite the urgency of the issue. The report points to the reliance on fossil fuels to power AI growth agendas in the two largest economies as a primary factor behind the surge in emissions.
Datacentres house the servers and infrastructure that process and store the vast amounts of data needed to train and run AI models. The energy consumption of these facilities has been rising steadily, but the rapid scaling of AI applications has accelerated demand in recent years. The IEA’s forecast suggests this trend will continue, with the environmental impact intensifying as AI becomes more embedded in business and consumer technologies.
The agency’s data shows global electricity use by datacentres will double by 2030. Three-quarters of that growth will come from AI-related activities, reflecting how the sector’s expansion is concentrated in computationally intensive tasks such as machine learning and large-scale data analysis. The US and China, which together dominate the AI market, are expected to contribute most to the increased energy use.
The IEA’s warning raises questions about the sustainability of current AI development paths, especially given the continued dependence on fossil fuel-generated electricity in many regions. While some datacentres are powered by renewable energy, the overall global trend still leans heavily on carbon-intensive sources.
This year’s conference has seen discussions around energy and climate policies, but the draft text due today reportedly lacks optimism. The rise in emissions from datacentres contrasts with efforts to reduce carbon footprints across other sectors, highlighting the complex challenges posed by emerging technologies.
The IEA’s forecast does not provide a breakdown of emissions by country beyond noting the US and China’s outsized roles. It also does not specify which fossil fuels or energy mixes contribute most to the increase. The agency’s focus is on the overall growth in demand and its environmental consequences.
The report arrives as AI technologies become increasingly central to global economies and societies. Their energy requirements are drawing more attention from policymakers, environmental groups, and industry leaders who face pressure to balance innovation with sustainability.
The next four years will test whether the datacentre industry and governments can shift towards cleaner energy sources to mitigate the expected rise in emissions. The IEA’s figures suggest this challenge is urgent, with the AI boom poised to reshape electricity consumption patterns worldwide.
According to The Guardian.







