AI Expansion Could Triple Data Center Water Use by 2030

Rystad Energy estimates that global data center water consumption could nearly triple to 644 billion liters per year by 2030, driven primarily by the expansion of artificial intelligence. Direct cooling water use stood at 222 billion liters in 2025, though the adoption of water-efficient technologies and active mitigation could lower the 2030 demand to 388 billion liters.

The total water footprint varies based on cooling architecture and the water intensity of electricity supplies. In the United States, indirect consumption may exceed direct use by more than double. Minh Khoi Le noted that forecasting these impacts is difficult because consumption varies by climate, cooling technology, and whether measurements track water withdrawn or actually consumed.

In response, hyperscalers including Amazon Web Services, Google, Microsoft, and Meta have pledged to become water positive by 2030. Governments are also implementing restrictions; the EU Commission is planning energy efficiency standards, and Singapore has introduced a Green Data Center Roadmap. Texas has imposed a moratorium on new data center approvals pending resource audits, as high-stress regions in Texas and India remain most vulnerable to increased demand.