Campaigners are warning that the rapid growth of AI could trigger a new surge in PFAS production, which are also commonly referred to as 'forever chemicals'. The demand is due to the physical infrastructure of AI, which includes the need for more chip fabs, more advanced manufacturing, and higher density data centres. The concern for insurers is that PFAS losses can take years to emerge, often lead to complex litigation, are costly, and they can trigger claims under several different policy lines at the same time.

The AI–PFAS link: where the demand comes from

The Guardian cites ChemSec’s finding that nearly all of the 10 biggest PFAS manufacturers have plans to increase production, with some producers stating “advanced data centres and AI hardware” as the factors for the growth.

Two parts of the AI supply chain that are the most relevant:

• Making the chips: PFAS can be used during chip production and in some specialist components used inside fabrication plants. Industry experts say these materials are hard to replace quickly because chipmaking needs extremely clean, tightly controlled conditions. 
• Cooling the data centres: As AI servers generate more heat, more sites are moving to liquid cooling. Some high performance cooling systems have historically relied on fluorinated engineered fluids, which campaigners say could increase demand for PFAS type chemicals. 

Implications on insurers

PFAS exposure can lead to claims under several routes, including:

• Environmental clean-up: costs to investigate, remove or treat PFAS in soil or water (including groundwater/surface water), plus issues with disposal and contaminated waste streams.
• Products and injury claims: allegations that PFAS in products or emissions causes harm, potentially developing into large group claims or mass tort litigation.
• D&O / securities: claims linked to how a company discloses and manages PFAS risk (e.g., alleged misstatements, inadequate reserves, or regulatory enforcement becoming financially material).
• Contract and professional disputes: disputes over who is responsible for contamination, warranties/representations in the supply chain, clean up obligations, and indemnities.

How the market is reacting

Insurers and brokers have taken the approach of increasing the restrictions of coverage, and PFAS-specific exclusions are becoming more common. Marsh for example, notes the rising use of PFAS exclusions as part of broader coverage tightening. 

In the London market, the Lloyd’s Market Association has also updated PFAS exclusion wordings (LMA5595/LMA5596), highlighting that the PFAS language is now an underwriting and wording issue, especially for casualty and environmental exposures. 

For insurers, the difficulty is determining how older policies, especially their pollution exclusions and key definitions, respond when PFAS is alleged in different ways, such as environmental contamination, a defective product, misleading consumer statements, or bodily injury.

What underwriters and claims teams should do now 

As AI-related PFAS demand grows, these are the practical steps insurers are increasingly taking:

• Exposure mapping: Identify which insureds have PFAS related activities. For example, what PFAS they use, where it’s stored, how waste is generated and disposed of, and any potential release points (e.g., cooling fluids, firefighting foams, wastewater discharges).
• Controls and governance questions: Look for evidence of strong risk management, such as closed loop systems, leak detection, supplier take back arrangements, clear disposal and destruction routes, and documented compliance with new and emerging standards.
• Wording clarity: Decide whether to exclude PFAS, sublimit it, or carve back coverage for specific 'essential uses'.