Article Masthead

  • Category: Market Intelligence
  • Topic: PFAS Supply Chain Expansion
  • Region: Europe
  • Signal Type: Market and Risk Trend
  • Published Date: 2026-09-14
  • Original Source Date: 2026-09-13

01 What Happened

ChemSec’s updated review of the world’s leading PFAS producers, published on September 13, shows that most major producers are still expanding capacity despite litigation, regulation and increasing public scrutiny. According to the report, the expansion is being driven by three main demand clusters: AI and data-centre infrastructure, semiconductor manufacturing, and lithium-ion battery materials.

ChemSec says producers are no longer framing fluoropolymer growth only around traditional industrial demand. Instead, they are explicitly linking investment to “the AI revolution”, microchip fabrication and advanced cooling applications. The report cites examples such as Daikin planning major fluoropolymer capacity growth tied to semiconductor demand, while other producers continue positioning PFAS-containing materials as necessary inputs for high-performance electronics and thermal-management applications.

The report also notes that the sector is not moving in a single direction. Some companies, including 3M and BASF, have announced PFAS exit decisions or phase-outs, while others continue expanding production or investing in adjacent product lines. ChemSec’s central warning is that partial restrictions may be outpaced by new supply unless policymakers adopt broader controls.

For the water sector, the key news is not simply that PFAS production remains controversial. It is that AI-related infrastructure growth may be increasing demand for fluorinated materials upstream of the data-centre or semiconductor site itself.

02 Key Takeaways

01 Most major PFAS producers are still expanding capacity, and ChemSec identifies AI infrastructure, semiconductor manufacturing and battery materials as key growth drivers.

02 The signal is upstream: some of the environmental burden associated with AI infrastructure may arise in fluorochemical production long before a data centre or chip plant begins operation.

03 The market is diverging, with some companies pursuing phase-outs while others continue to invest in fluoropolymer and related PFAS product capacity.

03 Why It Matters

The water implication is easy to miss if attention stays only on the end-use facility. Over the past year, discussion around AI infrastructure has focused heavily on site-level questions such as cooling demand, wastewater discharge, reclaimed-water use and power availability. Those are real issues, but they do not capture the full water-risk footprint of the supply chain.

If data-centre cooling equipment, semiconductor tools or battery materials rely on PFAS-containing chemistries, then part of the environmental burden may be displaced upstream into fluorochemical manufacturing. That shifts the question from “how much water does the facility consume?” to a broader one: “what contaminants, liabilities and residual streams are created across the value chain that supports this infrastructure?”

This matters because PFAS is not a standard industrial input. It is a class of highly persistent compounds associated with complex treatment, long-term liability and growing regulatory pressure. A cooling or electronics application may appear efficient at the point of use while still being linked to difficult wastewater and waste-management problems at the manufacturing stage.

In other words, a lower on-site water footprint does not automatically mean a lower overall water-risk footprint. For ATLAS readers, that is the more strategic signal.

04 ATLAS Engineering View

From an ATLAS engineering perspective, this development reinforces the need to evaluate industrial water risk beyond the fence line.

For data centres, the direct engineering discussion often revolves around evaporative cooling, liquid cooling, cycles of concentration, blowdown chemistry and reuse opportunities. For semiconductor manufacturing, it centres on ultrapure water production, rinse-water recovery and wastewater segregation. Those topics remain essential. But if the supporting materials and components depend on expanding PFAS production, then water stewardship can no longer be assessed only at the operating site.

The practical implication is not that every fluorinated material can be immediately replaced. In some advanced applications, substitutes may still be limited or performance trade-offs may be significant. However, procurement and design teams should increasingly ask where PFAS-containing materials are used, what alternatives exist, and whether a process efficiency gain at one point in the chain creates a larger persistence and treatment burden elsewhere.

The broader trend is that water intelligence is becoming supply-chain intelligence. For utilities, industrial end users and technology providers, the relevant question is no longer only whether AI infrastructure uses water efficiently. It is whether the infrastructure build-out is transferring water-quality risk into harder-to-manage chemical manufacturing pathways.

05 Sources

Image Source and Usage Notice

ATLAS engineering concept illustration showing how AI infrastructure demand can propagate upstream into fluorochemical/PFAS production and related industrial-water risk. It is a conceptual editorial visualization, not an owner-issued process drawing or photograph of a specific PFAS facility.

  • Credit: ATLAS WTS original engineering concept illustration.