The real money in AI chips is in these stocks - and you've probably never heard of them

By Charlie Garcia

Ultrapure water is scarce - and Big Tech has to have it for AI and semiconductors

Money is going to a few specialized companies that clean, move and monitor ultrapure water used in semiconductor manufacturing.

The companies purifying water look like better bets than the high-tech customers who use it.

Advanced chip-fabrication plants - fabs, in industry shorthand - can't run without ultrapure water. A handful of Japanese and U.S. companies control it - and they don't trade like meme stocks.

The U.S., Japan, Taiwan, the E.U., South Korea, China and India are shoveling hundreds of billions into chip plants and "semiconductor independence" - as if national security came in fab-shaped boxes. Politicians love hard hats and ribbon-cuttings. On TV it looks like progress.

What nobody mentions at the podium: The bottleneck isn't fancy lithography gear or geniuses in hoodies. It's water so obsessively pure it would make a Trappist monk feel dirty.

One advanced fab burns through roughly 10 million gallons of ultrapure water a day, about what a mid-sized American city uses. Scaled up, the chip industry drinks like a metropolis of several million people, except instead of flushing toilets they're rinsing transistors measured in atoms.?

While everyone argues about Nvidia (NVDA), the quiet money is going to a few specialized companies that clean, move and monitor that water.

You've never heard of them. That's the opportunity.

What makes water 'ultra' pure

Ultrapure water, or UPW if you want to sound smart at parties, is water stripped of almost everything.

Every dissolved ion. Every particle. Every organic molecule. Every microbe. Purity standards measured in parts per trillion. Not million. Not billion. Trillion.

Here's where it gets expensive.

The shift from older 28-nanometer chips to the bleeding edge, sub-5-nanometer and eventually 2-nanometer nodes now being built by TSMC (TSM), Samsung (KR:005930) and Intel (INTC) multiplies the rinse and cleaning steps per wafer. Smaller transistors are more vulnerable to contamination, so every generation demands more water, purified to tighter tolerances.

Leading-edge fabs use several times more water per wafer than older plants. Shrink the transistor, double the thirst. The accountants never saw that coming.

Semiconductors still account for roughly half of UPW demand. But biopharma is the fastest-growing customer. In pharma, the water doesn't just clean the equipment. It ends up in the product.

That IV drip better be pure.

Japan: The unlikely water superpower

The market structure is surprisingly concentrated. And Japan sits at the center of it.

This will annoy people who think everything important happens in Silicon Valley. Good.

Kurita Water Industries (JP:6370) (KTWIY) in Tokyo is one of the few firms that can design, build and operate UPW systems for leading fabs. Kurita has been shifting from selling equipment to "water as a service" contracts. In short, recurring revenue - two of the most beautiful words in investing.

Kurita is levered to Japan's semiconductor revival, including TSMC's Kumamoto fabs, plus ongoing spending by chipmakers across Taiwan and Korea.

For investors, Kurita is the sleep-better-at-night play. Diversified customers. Recurring revenue. Exposure to secular demand rather than betting on any single plant. Boring. Profitable.

Organo: The TSMC shadow: If Kurita is the steady compounder, Organo Corp (JP:6368) is the TSMC shadow play. It's less familiar to U.S. investors. That's because U.S. investors think Japan stopped mattering in 1989.

They're wrong. Organo has been a key ultrapure water supplier to TSMC for years. When TSMC builds a new fab in Japan or the U.S., Organo follows.

Like a very profitable remora attached to a very large shark. Organo's earnings have risen with TSMC's AI-driven capex. The stock has re-rated accordingly. Yet it still trades at a discount to its U.S. water and industrial peers.

If you believe TSMC remains the indispensable foundry for advanced chips, it's hard to ignore the company supplying the water those fabs cannot run without.

Nomura Micro Science: For gamblers with a thesis: Nomura Micro Science (JP:6254) offers a higher-beta expression of the same idea.

Translation: More upside, more volatility, more antacid. It's a much smaller Japanese specialist. Revenue heavily concentrated in microelectronics water systems, including projects for South Korean memory makers and overseas fabs.

After a run of record results, management guided for a near-term slowdown as the order book normalizes. The shares now trade at a single-digit earnings multiple despite secular growth in advanced-node capacity.

For investors willing to stomach cyclicality in semiconductor capex, Nomura is a leveraged bet on the timing and geography of the next build-out cycle.

U.S. stocks to consider

This theme isn't confined to Japan. Ecolab (ECL) is better known for hygiene and industrial water. It's the company that keeps hotel swimming pools from turning green. But its $1.8 billion acquisition of Ovivo's electronics-water business last August bought Ecolab a front-row seat in semiconductor UPW.

Ovivo brings customers including advanced fabs in the U.S., Europe and Asia. For now, UPW is a modest share of Ecolab's diversified revenue. But it's one of the more attractive growth legs in a company investors otherwise treat as a defensive compounder. The tortoise with a semiconductor kicker.

Xylem (XYL) approaches the problem differently. Less plant integration, more technology and data. Its acquisition of Evoqua in January 2023 broadened its high-purity water footprint. But the company's strength remains in pumps, sensors, analytics and cloud platforms that measure and adjust water quality in real time.

As fabs automate and adopt predictive maintenance, Xylem sells the instrumentation and software. They win regardless of who gets the engineering, procurement and construction contracts. They're Switzerland in the water wars.

DuPont (DD), meanwhile, sells the consumables that make UPW systems work: reverse-osmosis membranes, ion-exchange resins. Its brands include FilmTec and AmberTec, which nobody outside the industry has heard of and everybody who's inside depends on. Think razors, not razor handles.

Water Solutions represents about 12%-13% of DuPont's revenue. Not dominant, but not trivial. A larger Electronics & Industrial division sells materials like CMP pads into the same chip fabs, so DuPont hits semiconductor manufacturing from multiple angles.

Here's where it gets complicated, because this is corporate America and nothing can stay simple. DuPont is breaking itself apart. The electronics business spins off into something called Qnity. Water and other specialties stay behind in the rump DuPont.

Once the bankers finish collecting their fees and the org charts stop moving, the UPW contribution to what remains will become easier to see. Until then, you're buying a company mid-surgery and hoping the patient survives with the organs you wanted.

India's a wild card

Emerging markets are starting to matter to the UPW business.

VA Tech Wabag (IN:533269), based in India, has historically focused on municipal and industrial water treatment. Now it's moving up the value chain. It's securing orders for high-spec systems for solar-cell manufacturing, positioning for India's nascent semiconductor push.

New fabs under India's industrial policies will need modern water purification and recycling. Wabag's local presence makes it a natural partner despite the usual execution and policy risks that come with Indian infrastructure. For investors, Wabag expresses the UPW theme tied to India's industrialization rather than East Asia.

Why water ETFs won't do it for you

Water ETFs won't get you much UPW exposure. For example, Invesco Global Water ETF PIO holds four of these companies (Ecolab, Xylem, Kurita and Organo) - recently totaling roughly 16% of the fund. In the Global X Clean Water ETF AQWA, Kurita, Organo and Nomura Micro Science combined make up less than 4% of the fund. If you want the fabs' lifelines, you have to buy them directly.

How to think about your choices

These water companies sit at different points on business quality versus stock upside.

Durability: Kurita, Ecolab, and Xylem look the most bulletproof. They bring scale, balance-sheet strength and recurring revenue.

TSMC leverage: Organo is smaller but anchored to the most important foundry on earth.

Torque: Nomura and Wabag offer greater upside at the price of more volatility.

Nomura and Organo are most levered to leading-edge fabs. Wabag is tied to India's build-out. Kurita and Xylem are steady compounders. Ecolab and DuPont are safer conglomerate plays where UPW matters to the business but doesn't dominate it.

What could go wrong

There are still plenty of ways this investment theme can go sideways. Ultrapure-water projects are big, one-off beasts that everybody and his brother wants to build, so even the smart operators can watch their margins get mugged the minute capital spending gets a head cold.

If chip or biopharma outfits slam the brakes on new plants, or politicians decide to "review" fabs and exports, the order books won't crash so much as quietly stop returning your calls.

Then there's conservation, which is a virtue until someone sends you the bill. Recycling and tighter standards are supposed to save water, and they do, but only by bolting on more layers of treatment, sensors and consumables than a NASA life-support system.

You end up needing more gear to use less water. It's the rare public-policy miracle where everyone gets to feel morally superior while buying more expensive equipment.

The bottom line

(MORE TO FOLLOW) Dow Jones Newswires

12-09-25 2004ET

Copyright (c) 2025 Dow Jones & Company, Inc.