Artifi­cial Intel­li­gence: Why the AI Boom Is Trans­forming the Water Sector, Too

The AI boom is driving up demand not only for high-perfor­mance chips and data centers, but also for water techno­lo­gies. From ultra­pure water in semicon­ductor produc­tion to efficient cooling systems, new struc­tural oppor­tu­ni­ties are emerging along the water value chain.
AI-Stocks-Pull-Ahead-1164x840

AI Needs Water

Artifi­cial intel­li­gence is considered one of the most important drivers of growth in the coming years. The focus is usually on large techno­logy compa­nies and the massive expan­sion of data centers. Accor­dingly, the AI sector was respon­sible for a large portion of global stock market gains in the second quarter of 2026. But behind this trend lies not only a new techno­logy, but also an exten­sive physical infras­truc­ture and a resource that is often undere­sti­mated: water.

Water is needed for everything from the manufac­ture of high-perfor­mance chips to the construc­tion and opera­tion of data centers. The AI boom there­fore has an impact far beyond the techno­logy sector and is creating new demands on water treat­ment, cooling, and infras­truc­ture.

Source of chart: Tareno View, July 2026

No Chips, No Artifi­cial Intel­li­gence

The develo­p­ment of powerful AI models would not be possible without specia­lized high-perfor­mance chips. They handle the complex compu­ta­tional opera­tions on which modern AI appli­ca­tions are based.

Water plays a central role even in their manufac­tu­ring, as their produc­tion is very water-inten­sive. TSMC, one of the world’s leading chip manufac­tu­rers, consumed appro­xi­m­ately 151 million cubic meters of water in 2025. That amounts to more than 400 million liters per day. As chips become more complex, demand increases because more produc­tion and cleaning steps are required.

A large portion of this water is required as ultra­pure water, which is almost comple­tely free of minerals, particles, and other impuri­ties. This is where specia­lized water techno­logy compa­nies come into play. Kurita Water Industries, for example, develops and operates systems to supply the semicon­ductor industry with ultra­pure water. The Japanese company Organo Corpo­ra­tion also provides ultra­pure water systems and other water treat­ment solutions for the semicon­ductor and electro­nics industries.

But simply manufac­tu­ring the chips isn’t enough. To harness their enormous compu­ting power, additional physical infras­truc­ture is needed: data centers.

Chip-Herstellung--1279x1067

Why Data Centers Need Water

Data centers process large amounts of data and train and run AI models. As artifi­cial intel­li­gence becomes more widespread, the demand for high-perfor­mance data center infras­truc­ture is there­fore incre­a­sing. The high compu­ting power of the chips used generates signi­fi­cant heat, which must be reliably dissi­pated. Water-based systems and liquid cooling are playing an incre­a­singly important role in this regard.

Strong Growth Expected for Liquid Cooling in Data Centers

Source: Grand View Research, Kepler Cheuvreux; Chart by Tareno AG

This is where various compa­nies in the water sector come into play. Xylem offers solutions throug­hout the cooling water cycle—from pumps and filtra­tion to water treat­ment and digital monito­ring. Gorman-Rupp supplies pump systems used, among other things, for liquid cooling directly at server racks, in HVAC cooling systems, and in fire protec­tion. Watts Water Techno­lo­gies, in turn, provides valves, filtra­tion solutions, and other compon­ents that enable water to be controlled, treated, and effici­ently recir­culated in closed cooling systems.

As water-based cooling solutions become incre­a­singly important, there is growing pressure to use this resource as effici­ently as possible. This chall­enge is further compounded by the fact that many data centers are being built in regions with limited water availa­bi­lity. When selec­ting a location, other factors—such as power supply, grid connec­ti­vity, or land availability—often take prece­dence. This makes solutions that effici­ently treat and reuse water and keep it in closed loops for as long as possible all the more important.

What the AI Boom Means for Investors in the Water Sector

For investors, this develo­p­ment demon­strates once again that the economic impact of a long-term growth trend is not limited to its most obvious benefi­ci­a­ries. The expan­sion of artifi­cial intel­li­gence also creates demand among compa­nies whose techno­lo­gies are needed to build and operate the neces­sary infras­truc­ture.

In the water sector, this additional demand coincides with an already high struc­tural need for invest­ment. In many places, existing networks must be moder­nized, water losses reduced, and more water treated and reused.

Artifi­cial intel­li­gence is thus not the funda­mental growth thesis of the water sector, but rather an additional struc­tural driver that further reinforces the existing invest­ment need.

How Tareno Analyzes Oppor­tu­ni­ties in the Water Sector

For over 20 years, Tareno has been closely monito­ring compa­nies in the global water sector. The team at the Tareno Global Water Solutions Fund analyzes compa­nies across the entire water value chain and focuses on long-term develo­p­ments in the areas of water infras­truc­ture, water treat­ment, and efficient water use. New techno­lo­gical develo­p­ments and their impact on water demand are also factored into the analysis of the long-term oppor­tu­ni­ties and risks for indivi­dual compa­nies.

Subscribe to monthly updates on the Tareno Water Fund

Author

Katja von Krannichfeldt
Katja Von Krannichfeldt
Fund Manager & ESG Analyst

Note:

This publi­ca­tion is a promo­tional commu­ni­ca­tion. It was prepared for marke­ting and infor­ma­tional purposes and does not consti­tute an offer or a solici­ta­tion to subscribe to, buy, or sell shares of this invest­ment fund. It does not consti­tute invest­ment advice. Only the current fund documents (in parti­cular the prospectus and KID) are autho­ri­ta­tive. Past perfor­mance is not a reliable indicator of future results.