Caig: AI-guided water availability tracking and twin systems for infrastructure resilience

The National Science Foundation Office of Integrative Activities awarded the University of Alabama $1,214,308 on October 1, 2025, under the Geosciences program (CFDA 47.050) to develop AI-driven hydrologic models and digital twin systems for identifying water resources to support expanding AI infrastructure cooling demands across the United States.

The project will deliver three primary technical components. An AI-driven hydrologic model will analyze geospatial data across the contiguous United States to identify regions with adequate water resources for AI infrastructure sites. Digital twins will be created for selected sites to enable scenario analysis and reveal potential impacts of natural hazards and water availability fluctuations. A fractional-calculus-based modeling framework will assess environmental impacts associated with data center operations, including wastewater and thermal pollution outputs. Outcomes include novel methods for large-scale water resource analysis, new AI algorithms tailored to geoscience applications, and practical guidance on environmental management strategies.

The project runs through September 30, 2028, with work performed in Tuscaloosa, Alabama. The University of Alabama's Office for Sponsored Programs serves as the prime recipient. The program supports three graduate students and includes development of new graduate courses integrating AI and geoscience.

The American Geosciences Institute received a sub-award dated July 27, 2026, under the AI-Driven Hydrologic Modeling of Water Resources for Expanding AI Infrastructure project component.

Abstract

The rapid expansion of artificial intelligence (AI) infrastructure across the united states presents challenges for water resource management. AI infrastructure requires a large water supply for cooling. This project will develop three new AI models for water resources. The models will identify where water resources exist that can reliably support the growth of AI infrastructure. The models will use both real-time data and what-if simulations. First, a large-scale AI model will provide insights for water availability to choose local sites. Second, digital twins will be created to reveal hazards that may disrupt local water supply. Finally, an AI method will help predict environmental impacts such as wastewater and thermal pollution. The broader impacts include practical guidance on water management and sustainability. New graduate courses integrating AI and geoscience will be taught, and the project will support three graduate students. The project?s primary scientific goal is to develop a multi-scale hydrologic modeling framework that integrates physics-informed AI and hierarchical digital twin technologies to inform water management for AI infrastructure development. The project consists of three main technical components for analyzing water resource sustainability and identifying optimal sites for AI infrastructure. First, an AI-driven hydrologic model will analyze geospatial data across the contiguous united states to identify regions with adequate water resources. Second, digital twins will be created for selected sites, enabling scenario analysis to ask, ?what if?? And understand the potential impacts of natural hazards and water availability fluctuations. Third, a fractional-calculus based modeling framework will be developed to assess environmental impacts associated with pollution resulting from data centers. The outcomes include novel methods for large-scale water resource analysis, new AI algorithms tailored specifically to geoscience applications, and practical guidance on environmental management strategies. The research team includes experts in hydrology, AI, and applied mathematics, supported by collaborations with national labs and water management agencies. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.