Can AI Fix America’s Leaky Water Systems?

Utilities test AI pilots from Xylem, Claros, and Oldcastle to modernize U.S. water infrastructure

10 Oct 2025

Across the country, water utilities are dipping their toes into artificial intelligence, hoping data can help rescue aging systems. From leak detection to predictive maintenance, a wave of pilot projects hints at how digital tools could make America’s water networks smarter and more efficient.

Xylem, a global name in water technology, has broadened its partnerships with filtration and membrane innovators, pairing real-time sensing with advanced treatment systems. Analysts see the move as a step toward integrated operations that could tighten efficiency across city networks without the need for costly overhauls.

In Minnesota, Claros Technologies is testing its high-flow treatment setup with industrial partner Daikin. The pilot aims to prove that PFAS, or “forever chemicals,” can be destroyed at scale, helping utilities meet tough new EPA standards.

Meanwhile, Oldcastle Infrastructure is running a leak detection trial with a Midwestern utility, using AI to analyze real-time data and locate hidden water losses. Early reports show promising reductions in both waste and maintenance costs.

These efforts are boosted by federal programs under the Bipartisan Infrastructure Law and the Inflation Reduction Act, which fund digital monitoring, cybersecurity, and energy upgrades. The support gives utilities a financial cushion to test new technologies without gambling on unproven systems.

Yet challenges remain. Data governance, workforce readiness, and the tangle of legacy infrastructure continue to slow the digital shift. Even so, investment in smart water solutions is rising steadily as utilities experiment with what AI can do at scale.

For now, the transformation is gradual but unmistakable. One pilot at a time, America’s water systems are learning to think for themselves.