Kemira, a Finnish water treatment chemicals company, and CuspAI, a Cambridge-based AI materials science startup, have used generative artificial intelligence to design new materials capable of removing PFAS compounds from drinking and industrial water at trace concentrations, compressing what would typically take years of research into six months.
The collaboration produced more than 5,000 novel material designs targeting three priority PFAS molecules: GenX, PFBS and PFOS, from a search across approximately 300 trillion possible structures. Around 20 candidates have been selected to advance to development and testing.
PFAS, or per- and polyfluoroalkyl substances, are a family of synthetic chemicals used across industries since the 1940s. They are known as "forever chemicals" because they do not break down naturally in the environment and accumulate in the human body. Regulators on both sides of the Atlantic have tightened limits on their presence in drinking water. The US Environmental Protection Agency introduced maximum contaminant levels at the parts-per-trillion scale in 2024, while the EU has updated its Drinking Water Directive accordingly.
The collaboration produced more than 5,000 novel material designs targeting three priority PFAS molecules
The dominant remediation technology today is granular activated carbon. Kemira, which is already active in the activated carbon regeneration market, initiated the collaboration to explore whether AI-driven design could yield more selective and longer-lasting alternatives. The approach focused on metal-organic frameworks, nanoporous crystalline materials whose structure can be precisely tuned for specific filtration tasks, with CuspAI's platform generating entirely new structures from scratch against industrial performance criteria including water stability, environmental compatibility and cost-effective synthesis. The project also uncovered new functional group chemistries, the partners say, that could have broader applications in adsorption products.
The companies describe it as the first commercial partnership to apply generative AI end-to-end to PFAS materials design, moving beyond screening known compounds to designing new ones at scale.
Chad Edwards, chief executive of CuspAI, said the partnership had delivered on "one of the most pressing environmental problems of our time" and that further projects were already being scoped under a framework agreement. Kemira's president Antti Salminen described the outcome as "a credible path toward a next-generation PFAS remediation product."
