AI Designs Materials to Remove Forever Chemicals in 6 Months

Scientists just cracked a major code in the fight against toxic "forever chemicals" polluting our water, and they did it in record time using artificial intelligence.

Kemira, a global water treatment company, partnered with AI firm CuspAI to create entirely new materials capable of removing PFAS contaminants from drinking water. These dangerous chemicals, nicknamed "forever chemicals" because they never break down naturally, have contaminated water supplies worldwide and have been linked to cancer and other serious health problems.

The AI platform examined nearly 300 trillion possible material structures to find the perfect solution. Within six months, it generated over 5,000 promising designs specifically targeting three of the most harmful PFAS molecules: GenX, PFBS, and PFOS.

The teams narrowed these candidates down to nearly 20 high-priority materials now moving into testing and validation. What makes this achievement remarkable is the timeline. Traditional materials discovery typically takes years or even decades.

This marks the first time a commercial partnership has used generative AI from start to finish to create completely new molecules designed for cleaning up PFAS pollution. Previous efforts relied on AI to screen existing materials, but this approach built new molecular structures from scratch based on what industrial water treatment actually needs.

The materials aren't just theoretically promising. They're designed to work at extremely low concentrations where PFAS typically appears in contaminated water, while remaining stable, sustainable, and affordable enough to manufacture at industrial scale.

Why This Inspires

This collaboration shows how cutting-edge technology can tackle environmental crises affecting millions of people. PFAS contamination has become one of the world's most urgent water safety challenges, present in everything from firefighting foam to food packaging and now detected in drinking water across countless communities.

By compressing years of research into months, AI-driven materials design could accelerate solutions not just for PFAS but for other persistent environmental toxins. The partnership has already opened discussions for additional water treatment innovation projects.

The promising candidates now move into real-world validation testing, bringing us closer to affordable, scalable solutions that could protect drinking water for communities worldwide.