Hong Kong Data Centers Exceed Global Average in Carbon Footprint, UN Study Reveals

A comprehensive study released by the United Nations University Institute for Water, Environment and Health on June 8, 2026, has brought to light the substantial environmental impact of data centers, particularly in Hong Kong. The report reveals that Hong Kong's data centers operate with a carbon footprint that is 43% above the global average, positioning it among the most carbon-intensive grids alongside Indonesia and India. This elevated footprint is largely attributed to the city's energy mix, where fossil fuels account for a dominant 67% of the power supply. The study, led by environmental scientist Kaveh Madani, extends its analysis beyond just carbon emissions, quantifying the water and land footprints associated with the electricity consumed by AI systems. It emphasizes that every kilowatt-hour used for AI training, and especially the continuous operation of AI models (which accounts for 80% of total energy use), carries significant water and land resource implications. These include the water needed for cooling heat-intensive data centers, generating power, and the land required for energy infrastructure and fuel extraction. Without proper management, these processes risk depleting vital natural resources. Globally, data centers generated 189 million tonnes of carbon emissions last year, an amount that would necessitate the growth of 3.2 billion tree seedlings over a decade to offset. The report identifies Hong Kong as one of the world's top 20 data center hubs, a list headed by the United States, Germany, the United Kingdom, and mainland China. Despite its high carbon intensity, Hong Kong scores below the global average for water and land footprints in electricity production, mainly because its energy mix uses less hydropower or other renewable energy sources that typically have larger water and land demands. In response to these findings, the UN study advocates for a responsible approach to AI development and deployment, proactively addressing its unintended environmental consequences to ensure sustainability and equity. Key recommendations include the adoption of disclosure standards for AI's environmental impact, covering carbon, water, and land footprints. Furthermore, the report advises investors to thoroughly vet projects based on these environmental considerations and encourages companies to factor in these footprints when making decisions about where to establish new data centers. This proactive stance is crucial for mitigating the ecological burden of the rapidly expanding digital infrastructure.