The state of ultrapure water: Insights from UltraFacility

For decades, the goal of water systems for microelectronics was straightforward: achieve the highest purity possible. That objective remains essential. But today, purity alone does not define performance. 

As production increases to meet the AI boom and chip design become more complex, semiconductor manufacturers need ultrapure water systems that operate with stability. Increasing internal reuse, variable source waters and tighter operating margins mean that the question is no longer simply, “How clean is the water?” It is, “Can the system deliver that level of purity reliably, every day, at full production scale?” 

At UltraFacility 2025, a leading forum for semiconductor facility infrastructure, engineers and operations leaders focused on this shift. Water systems are no longer facility infrastructure. Their stability and adaptability directly affect yield, uptime, cost control, and the ability to scale production with confidence.

Insights from Xylem’s microelectronics leader 

Below, Alan Knapp, Senior Director, Business Development – Microelectronics at Xylem provides his takeaways from UltraFacility 2025. Alan and his team help semiconductor customers address ultrapure water, wastewater and sustainability challenges. With more than four decades in industrial water, he works directly with manufacturers navigating increasingly complex purity and reuse requirements. 

You are a long-time UltraFacility attendee. How has the focus of the conference and the microelectronics industry evolved over time? 

When I first attended UltraFacility, the conversation centered on pushing purity limits — removing more contaminants, measuring lower detection limits and optimizing individual treatment technologies with a focus on reducing water usage in the process. That still matters, especially as chip designs shrink and tolerance for variability drops, while the amount of process steps has increased, leading to a higher UPW demand.

What has changed is the broader context. Production volumes are growing rapidly, driven in part by AI and advanced computing demand. At the same time, fabs are increasing internal water reuse to manage cost and utility supply risk. 

The focus has shifted from achieving peak purity in isolation to maintaining system-wide stability under real operating conditions. The question today is: How do we keep water consistently clean while scaling production and maximizing reuse — without putting yield at risk?

How are changing source waters creating new reliability risks for UPW systems? 

Source water variability is becoming a practical design consideration. In some regions, fabs alternate between different municipal supplies — such as surface water and groundwater — which can have very different chemistry. In Gresham, Oregon, facilities may switch between well water and surface water. In parts of the Bay Area, manufacturers rely on multiple water sources with distinct mineral and organic profiles. 

Even modest chemistry shifts can impact membranes, ion exchange resins and downstream polishing systems. 

That said, in many cases the greater variability comes from inside the facility. As facilities recycle more process water back into the UPW system, residual contaminants become more concentrated. Each increase in reuse tightens the operating margin. 

Small upstream changes, whether from blended source water or an internal recycled stream, can ripple through the treatment train. Designing for variability, not ideal baseline conditions, is critical. That means integrating flexible pretreatment, real-time monitoring and adaptive control strategies that maintain stable performance as inputs and recovery rates evolve. 

How is water reuse shaping UPW system design? 

Water reuse is both a sustainability initiative and a risk management strategy. It reduces freshwater demand and mitigates supply risk, but it increases system complexity. 

Producing ultrapure water always generates streams that contain most of the removed impurities. As facilities push recovery rates higher, those streams become smaller but more concentrated. If not carefully managed, that concentration can stress membranes, resins and technologies used for recovery. 

Effective reuse requires a systems-level approach — understanding how each stage of the treatment train interacts and ensuring that higher recovery does not compromise reliability or yield. 

What does AI-driven growth mean for ultrapure water? 

The growth of AI and digital infrastructure is increasing semiconductor demand. While data centers have different water requirements, they amplify pressure on water and cooling systems, especially in water-stressed regions. Data centers are also looking at effective liquid cooling methods to reduce water consumption. 

For fabs, this makes water planning more strategic. Facilities cannot assume stable, abundant supply. Reuse capability, resilience and operational flexibility are now factors in expansion decisions. As AI drives production growth, the importance of reliable, adaptable UPW systems increases alongside it. 

 What differentiates successful ultrapure water partners today? 

There is no one-size-fits-all answer. Every facility has unique source water, production goals and recovery targets. 

Successful partners understand how the entire system works together. Ultrapure water is not one technology — it is a chain of technologies that must operate in balance. Experience matters because decisions made at one stage affect everything downstream. Water circularity is now the focus. 

What differentiates strong partners is the ability to apply that experience to real operating conditions. It is not just about installing equipment. It is about helping facilities design systems that remain stable as conditions evolve. 

How important is workforce development to the future of UPW systems? 

Workforce development is critical. UPW systems are increasingly complex, requiring expertise in chemistry, hydraulics, process control and system integration. 

At UltraFacility 2025, Xylem partnered with Global Water Intelligence (GWI) to host an engineering challenge for students focused on semiconductor water systems. Participants received dedicated mentoring, hands-on guidance and insight into real-world system design. Initiatives like this demonstrate that the future of ultrapure water relies not just on technology, but on passing knowledge and expertise to the next generation of engineers who will operate, optimize, and advance these critical systems.

From purity to performance: What comes next 

UltraFacility 2025 made one thing clear: UPW systems must do more than deliver clean water — they must adapt. Manufacturers need systems that adapt to variable inputs, support higher reuse rates and remain stable at production scale. 

Designing for resilience, flexibility and integrated performance is now as important as achieving low contaminant levels. In 2026 and beyond, ultrapure water will be defined not only by how clean it is — but by how consistently and confidently it supports growth.