TheChek News Flow

Power gets all the attention, but water is a growing issue.

Water, water everywhere, but…is too much going to artificial-intelligence data centers? That’s the challenge—and opportunity—facing industrial stocks that specialize in H2O.

AI is the growth engine of this stock market, driving everything from utility earnings to SpaceX’s planned $2 trillion initial public offering. That makes it important for investors to understand any potential AI bottlenecks. While power gets all the attention, water is a growing issue. More-powerful AI chips need water to cool them, and managing that increasingly scarce resource is now mission-critical for any hyperscaler that wants to maintain good public relations and be a reasonable steward of the environment.

The issue is only going to get more critical. Nvidia’s H100 chips are currently the most widely deployed AI chip. They can still be cooled by what are essentially big fans blowing air from a giant air conditioner. Such air conditioners, like home ACs, are essentially closed-loop systems, in which a refrigerant circulates within a sealed system. Newer chips use more power, necessitating new cooling solutions, including direct-to-chip cooling, where a plate is attached to the processor. It’s a bit like the way car engines are cooled, with coolant circulating through the equipment. Eventually, chips will need to be cooled by immersing the server in a liquid and with special evaporating liquids, though that’s still years down the road.

This will require lots of water. Morgan Stanley estimates that AI water use will grow to more than 1 trillion liters by 2028, or 400,000 Olympic-size swimming pools. That includes water for power generation, much of which gets recirculated, as well as for cooling and other purposes, so the ultimate amount may be less. Still, it will be up to industrial companies to build systems that can cool chips in efficient closed-loop systems, with as little waste as possible. Here are six stocks that should benefit.

Packaging power and cooling together is a competitive advantage—and Eaton, a hardware and software provider for data centers, is on its way to doing just that. In March, Eaton closed on its acquisition of Boyd Thermal, which provides both power and cooling for AI data centers. That makes the company a system provider, which gives it an edge over companies that provide only components, says Janus Henderson research analyst William Brothers. The deal also gives Eaton 500 more engineers specializing in cooling tech. With expected earnings growth and its recent valuation, shares could fetch about $470 in a year, up 19% from recent levels.

Like Eaton, Schneider Electric provides both electrical hardware and software for data centers. That is the result of its acquisition of 75% of Motivair in February 2025, which brought expertise in cooling distribution units and direct-to-chip cooling plates in house. The deal has been a tailwind for Schneider. RBC Capital analyst Mark Fielding forecast total sales growth of 9% a year through 2030, up from roughly 6% annually over the past three years. He rates shares Buy and has a $68.40 target for the U.S.-listed American depositary receipt, up 10% from recent levels.

Vertiv

VRT Holdings didn’t need to make an acquisition to merge power and cooling technologies—it already had them. The stock has returned 70% annualized over the past five years, but there’s no reason it can’t keep rising. Shares jumped 25% on Feb. 11 after fourth-quarter numbers, which showed orders had grown 252% year over year. Earnings estimates for 2026, now at $6 per share, rose 40% over that span. The stock trades at 46 times expected earnings over the next 12 months, but profits should grow by 34% annually over the next three years. And Wall Street tends to underestimate Vertiv’s earnings.

Modine Manufacturing isn’t a household name, but D.A. Davidson’s Matt Summerville calls it a “best in class” data-center play. Modine provides everything data centers need to cool chips: rear-door heat exchangers, AC units, chillers, and coolant distribution units. Overall sales grew 7% in fiscal 2025, driven by 28% growth in its Climate Solutions segment, helping earnings before interest, taxes, depreciation, and amortization grow 25%. That growth, which should continue, justifies its price/earnings ratio of 32 times earnings over the next 12 months. After divesting its auto parts business later this year, Modine will be a pure-play cooling company.

After heat is removed from chips, data centers still need systems to manage and cool water. That’s where Trane Technologies comes in. It builds coolant distribution units and air handlers, and is investing in immersion cooling, buying LiquidStack in March. Wall Street expects sales and earnings to grow by 8% and 13% a year on average, respectively, for the next three years. Above-average earnings growth is a big reason shares trade for 31 times earnings estimated over the next 12 months. Next-generation cooling could push profits higher. Bookings in Trane’s Applied Solutions business, which serves data centers, grew 120% in 2025’s fourth quarter.

Solstice Advanced Materials, which makes chemicals used in semiconductors, uranium hexafluoride for nuclear power, and refrigerants and coolants, was spun out of Honeywell International in October 2025. It serves a variety of industries, including data centers. Wall Street wasn’t a huge fan of the stock when it spun out. Analysts expected earnings to grow at just 15% annually at the time of the spin. Investors were quicker to realize Solstice’s potential. Shares are up 65% year to date, and analysts have come around too—they now see earnings growing at a 20% clip annually through 2028.

Outgoing Prime Minister Orban fostered strong ties, making Hungary a key European hub for Chinese manufacturing.

China invested heavily due to Hungary’s 30% to 40% cheaper labor, 9% corporate tax, and Orban’s political support, despite environmental and labor concerns.

Successor Peter Magyar seeks a balanced approach to China, welcoming investment but prioritizing environmental protection and added value.

Outgoing Hungarian Prime Minister Viktor Orban was famously Russia’s best friend in the European Union. He was China’s, too. That presents more thorny policy choices for his successor, Peter Magyar.

Beijing eclipsed Germany as Hungary’s most active foreign investor in recent years, turning the country of 9.6 million into its European bridgehead for advanced manufacturing. Fujian-based Contemporary Amperex Technology, or CATL, the world’s top electric vehicle battery maker, is slated to open a 7.3 billion euros ($8.6 billion) Hungarian plant this year. Leading EV producer BYD has committed €4 billion for its own factory. Telecom equipment giant Huawei Technologies established its European hub outside Budapest.

Part of the lure was economic. Hungarian labor remains 30% to 40% cheaper than Germany’s, says Zoltan Kiszelly, director of the Center for Political Analysis at the Szazadveg Foundation. Corporate taxes, at 9%, are the lowest in the EU save Ireland and Cyprus.

Orban added political sweeteners, vetoing or dissenting from numerous EU resolutions condemning Chinese behavior around human rights, Hong Kong or Taiwan. “The reason China chose Hungary over Poland or Czechia was the friendship factor,” says Agnes Szunomar, who heads the Institute of Global Studies at Corvinus University of Budapest.

Orban’s government may also have been soft on environmental and labor questions. Chinese workers building the BYD plant were subject to “internationally recognized indicators of forced labor,” alleges a recent report from U.S.-based NGO China Labor Watch. CATL’s neighbors-to-be are nervous about toxic waste, after revelations that Orban ignored reports of pollution from a Samsung SDI EV battery plant near Budapest.

All this comes as the EU hierarchy in Brussels pushes a “de-risking” strategy from China, though details are amorphous.

Magyar has mapped a balanced approach toward the Chinese golden goose. “I am very happy to travel to Beijing,” he told a press conference following his landslide April 12 victory. But “we will not accept foreign companies who create little added value and endanger Hungary’s land, air and water.”

Nixing the marquee investments by BYD and CATL looks vanishingly unlikely. The new premier should rather push for more EV supply chain and R&D investment in Hungary, Szunomar says. “Hungary is in an assembly trap with no tech spillover effects,” she says.

Huawei and its Chinese telecoms rival ZTE look like softer targets for a Magyar government as the EU moves toward banning them in 5G networks. “De-risking means the Chinese will be phased out and replaced by Ericsson, GE and Nokia,” Kiszelly predicts.

Magyar is also leaving himself wiggle room on whether and how to wean Hungary from Russian energy imports, which Orban maintained under an exemption from EU strictures. “Diversification needs to be strengthened, but this won’t happen overnight,” he told reporters on election eve.

The relationship with China could be more strategically important going forward. Each side has, in theory, a great deal to gain. Beijing has alienated most of the other European partners it courted through the so-called 17+1 initiative a decade ago, notes Dimitar Lilkov, a senior research officer at the Wilfried Martens Centre for European Studies. “Most of the investments fell through,” he says. “Chinese don’t like Western stuff like the rule of law.”

The flood of German investment that powered postcommunist Hungary meanwhile slowed to a trickle, making other sources essential. Magyar may need to tear out Orban’s domestic system by the roots. Ties with China require more delicate surgery.

The bet is that the company can regain its place at the semiconductor table—and that shares still have plenty of room to run.

Intel stock has been on a tear. Investors should consider buying it anyway.

Coming into Tuesday, Intel shares were on a nine-day winning streak, gaining 58% over that span and leaving shares up 220% over the past 12 months. Normally, Barron’s would shy away from recommending it. Our job is to identify stocks before they take off, not tell our readers to buy after shares have more than tripled. For a stock that has been as risky as Intel—its shares have been 1½ times as volatile as the S&P 500 index over the past year—getting it wrong will be painful.

Consider, however, where Intel is coming from. It’s starting from a distressed base—even after the recent gain, it lags behind the S&P 500 by 11 percentage points annualized over the past five years—with no support from Wall Street amid management turmoil. Now, with new leadership, a vastly improved product road map, and a solid strategy, Intel has a path to joining the sector elite, whose ranks include Taiwan Semiconductor Manufacturing, Broadcom, and Nvidia. While the stock dropped 2.1%, to $63.81, on Tuesday, it could hit $150, up 140%, in the coming years.

“There is hair on the story,” says Krishna Chintalapalli, portfolio manager of the Parnassus Value Equity and Value Select strategies. “If they get things right, the payoff is huge.”

Intel, as investors are painfully aware, has watched helplessly as other tech companies racked up huge gains. Its stock peaked at $75.87 in August 2000—near the top of the dot-com boom—and has never really come close since. While a pandemic boost for new laptops helped push shares as high as $69.29 in 2020, the stock hit a multiyear low in June 2025, below $18 a share and below book value. That 75% decline from its 2000 peak was the result of missing manufacturing transitions, such as the move to extreme ultraviolet semiconductor manufacturing, the shift to mobile computing, and the rise of graphic processing units, or GPUs, over central processing units, or CPUs.

Intel was simply left behind. In 2000, the company reported sales of about $34 billion and an operating profit of more than $10 billion. In 2025, sales were $53 billion and the company lost $2.2 billion, while Nvidia grew sales from $735 million to $215 billion, with operating profit climbing from $59 million to $137 billion. A string of five successive CEOs, including 80486 processor-designer Pat Gelsinger, couldn’t fix the problem.

Enter Lip-Bu Tan, a venture capitalist and turnaround expert who was responsible for the revival of Cadence Design Systems, which gained more than 3,200% over his 12-year tenure.

Tan took the reins in March 2025 and has moved quickly. He cut a bloated workforce by more than 20,000 and dramatically cut cash burn. Intel’s free cash flow was a cumulative negative $44 billion between 2022 and 2025—a $107 billion swing from the prior four years—as investment spending shot higher. Free cash flow was positive in the second half of last year and should improve in the coming years as capital spending normalizes and revenue grows.

Now Tan is focused on improving Intel’s manufacturing leadership and getting it back into the artificial-intelligence fight. Intel and Alphabet are working together on AI and cloud infrastructure, while Intel will help Elon Musk build and operate his “Terafab,” an AI chip-making joint venture between SpaceX and Tesla. Intel is also developing chips designed to run AI agents on laptops and AI accelerators, similar to Alphabet’s TPUs and Amazon.com’s Tranium chips.

In September, Nvidia invested $5 billion in Intel, which will make custom x86 server CPUs to integrate with Nvidia products. Nvidia’s GPUs are expensive, and Intel CPUs are a cost-effective way to support AI computing. “The demand for the x86 server CPU has gone through the roof at hyperscalers,” says Melius Research analyst Ben Reitzes. “It’s not gonna stop…the x86 became an AI chip.”

Still, much of this good news is reflected in the stock. Shares now trade for 95 times expected earnings over the next 12 months, surpassing their previous peak of 60 times reached in 2021, and fetching more than Nvidia, Taiwan Semi, Broadcom, and Advanced Micro Devices.

On the surface, Intel is anything but cheap. Earnings, however, are depressed—the company is expected to report 2026 earnings per share of roughly 50 cents in 2026, down from almost $5.50 in 2021. Intel’s gross profit margins in 2025 were below 40%, while Taiwan Semi’s were 55% and Nvidia’s were 75%.

There are reasons Intel’s margins are low. It’s paying its competitor Taiwan Semi to make an estimated 30% of its wafers as it builds its own fab capacity and invests in new products.

Intel’s manufacturing yields—the ratio of functional chips to the maximum possible—in its newest chip-making process are too low. Taiwan Semi’s yields are estimated to be as high as 90%, while Intel’s are estimated at about 70%. Improving yields isn’t a given, but as that process matures, there should be a new stream of earnings and free cash flow.

“Lip-Bu is great,” says Molly Pieroni, president of Yacktman Asset Management. “If anybody can pull this off, he can.”

Wall Street isn’t a believer. Earnings estimates for 2026 and 2027 have barely budged, and only about a fifth of analysts covering it have a Buy rating on the shares, well below the 55% for the average stock in the S&P 500. But analysts are starting to come around. Reitzes, for one, upgraded shares to Buy in January with a $75 price target, up 19% from recent levels, and others could follow. “If agentic AI and the importance of the semiconductor domestic supply chain are just in the early innings, the stock will grow in and grow into and surpass its valuation,” says Reitzes.

But even Reitzes’ target undersells what is possible. With a better product lineup and normal profit margins, Intel could easily earn $7 a share by 2029. At a typical semiconductor multiple of 22 times forward earnings, that would support a $150 price for Intel stock in a couple of years. Meanwhile, investors can look forward to rising earnings estimates and improving Wall Street sentiment. And even after Intel’s recent run, the company is worth just $320 billion, less than AMD’s $415 billion, despite 50% more sales. Nvidia, Taiwan Semi, and Broadcom are worth trillions. It certainly doesn’t hurt that Intel is the only U.S. chip maker—arguably an essential attribute due to security concerns.

For investors reluctant to go all in now, there’s always dollar-cost averaging —buying a little now and adding additional stakes in the weeks ahead. That way, if the stock keeps rising, at least they own some; if it drops, they can buy on the dip.

Just don’t miss it completely. The bet is that Intel can regain its place at the semiconductor table—and the stock still has plenty of room to run.

Chaos is the new normal, said the luxury brand's executive chair, as he shrugged off U.S. tariffs and said the group is exploring expansion into Saudi Arabia.

PARIS — Hermès International is leaning on its long-standing strategy of product development and geographic diversification as it navigates a more fragmented global luxury market shaped by geopolitical tensions, shifting wealth dynamics and uneven demand across regions.

Speaking at the company’s annual shareholder meeting, executive chairman Axel Dumas struck a confident tone about the group’s performance despite macroeconomic volatility and currency headwinds, pointing to resilient demand and the structural strength of its regional-focused business model.

“I will be very blunt — I am very happy with our first quarter,” Dumas said, noting sales were up 6 percent at constant currency even as a strong euro weighed on reported results and geopolitical disruption dragged down March sales in the Middle East.

“We had a slight impact in the first quarter due to geopolitics, but we are not responsible for exchange rates or geopolitics,” he said. “It remains very difficult to predict crises.”

Dumas framed the current environment through what he described as a structurally more unstable global order. “Crises are inevitable and unpredictable,” he said, referencing what he called a “VUCA” — volatile, uncertain, complex, ambiguous — world, where chaos has become the new norm, rather than an exception.

Despite that backdrop, he argued Hermès is structurally positioned to absorb shocks due to its balance across regions and categories, as well as its focus on cultivating local clientele.

“The Hermès model — balanced across métiers, geographies, with a strong focus on quality — seems to me a good strategy in such an unstable world,” he said. “So I remain optimistic about the fundamentals of Hermès and our ability to evolve in an anxious world.”

A key shift, Dumas said, lies in how luxury demand is now driven less by macroeconomic growth and more by changes in personal wealth linked to financial and property markets.

“The biggest transformation I have seen in recent years in the behavior of our clients .… Until the mid-2010s, the best indicator was GDP growth,” he said. “From the mid-2010s, the proxy that gives the strongest signal is real estate valuation and stock markets.

“People buy much more according to the evolution of their wealth than according to the evolution of national income,” he said.

That shift has been particularly visible in China, where weakness in real estate has weighed on consumption among parts of the population, he added. “The problems we can see today in China, for example, come first from a decline in real estate, which leads part of the population to spend less and save more,” Dumas said.

At the same time, he said geopolitical fragmentation is reshaping customer demand, with local consumption and “multilocal” purchasing patterns becoming more important.

“Because we are a small-volume brand, because we are exclusive, because this quality is not found elsewhere — we continue to see very strong interest across all geographic zones,” Dumas said.

He pointed in particular to the Middle East, where the current conflict has disrupted tourism flows but not eliminated demand.

“Where we see a risk area is the Middle East,” Dumas said. “We were growing strongly there before the war. If the high-spending, traveling clientele is affected, we too can be affected, though more moderately.”

He contrasted this with past crises, noting that during the pandemic, Hermès temporarily closed most of its stores, while the current disruption has been more uneven. And like the post-pandemic “revenge spending,” Dumas believes that pattern will hold.

“After a crisis passes, we often see a rebound in consumption — a desire to enjoy life again,” he said. “That is human nature.”

On pricing power and tariffs, Dumas downplayed their relative impact compared to structural cost pressures in Europe.

“Everyone talks to me about tariffs, but tariffs cost us much less than producing in France with the ‘exceptional contribution,’” he said, taking the first of a few swipes at the French government’s “temporary” tax applied to large corporate profits, which has been reintroduced several years in a row.

Dumas added that Hermès has increased prices in the U.S. to offset the cost of tariffs, which has not dented sales in the country. “Tariffs have been a small obstacle, but not the biggest obstacle in Hermès’ path,” he said.

A defining feature of Hermès’ model, he added, is its rejection of conventional marketing-heavy luxury strategies.

“I would say our first marketing tool is in the price of the product,” Dumas said. “We hope the product speaks for itself, and we therefore have lower marketing costs.”

He contrasted this with competitors that rely more heavily on advertising to support higher production volumes. “Some put the money elsewhere and then have to spend more on marketing to explain that [the product] is great.”

Dumas was also questioned on the company’s policy on counterfeits, “dupes” and the second-hand market. He referred to both reputational risks and cultural complexity, but downplayed their long-term impact on the brand.

Geographically, Hermès continues to expand selectively, prioritizing markets where local infrastructure and demand justify direct presence.

Regarding Saudi Arabia, where other luxury peers have been expanding rapidly, Dumas said they are searching for the right move now that Saudi Arabia’s foreign investment and commercial codes under its Vision 2030 plan have made it easier for international groups to operate more directly in the kingdom than under the previous, more licensing- and partnership-heavy rules.

“Now we can operate directly, so it is part of our thinking. We have a very strong Saudi clientele. Some clients travel to Europe or other stores in the region. There is strong demand there,” he said. “For now, we are negotiating. We are working on it, but we need to find the right partner locally.”

On India, Dumas described a market full of potential but structural constraints. “India is a very interesting country. We have a good Indian clientele, but they do not necessarily buy in India,” he said, citing high import duties, strong domestic artisanal competition and local style. “It is a formidable country where we will develop, at a pace that remains somewhat mysterious.”

Dumas struck a similar tone on Africa, where projects are under consideration, but indicated that any opening is still some way off.

“We have not yet found the right locations in terms of retail infrastructure and sufficient middle-class development,” he said. “We do have a strong African clientele who travels to our stores, but we usually establish ourselves in a country only when there is a sufficiently large local customer base.”

Intertek is going to be broken up: the question is by whom

A split should unlock value

Great minds think alike. Intertek, which carries out safety testing and certification, is weighing plans to splice off its less profitable energy and infrastructure unit, via a sale or demerger. It has rebuffed an offer from Swedish private equity firm EQT, which wants to buy the whole lot for £10.6bn, including debt, and carry out much the same slicing strategy.

It’s not possible to tell who had the lightbulb moment first. EQT launched its takeover bid on April 10; Intertek spelt out its plans a few days later but said discussions had been under way for some time. It also categorically rebuffed EQT’s advances.

Either way, a split should unlock value. For one thing, the two business lines are very different: each has its own labs and software. And, for another, the combination is not much loved by the market.

Before the news, Intertek’s enterprise value was just north of £7bn. Yet putting its testing and assurance business on 15 times this year’s estimated operating profit — in line with European rivals, which trade at somewhere between 13 and 17 times — means it, alone, would be worth £7.3bn. A more modest multiple of 10 for energy and infrastructure, whose margins are only half the size of the testing business, values that at £1.7bn, putting the combination at £9bn.

That’s a lot better than the shares were trading at last week. But, of course, EQT’s offer trumps that — as well as offering jam today to those investors who lack faith in Intertek’s ability to pull off such a transaction.

Indeed, the group’s record isn’t one of unalloyed success when it comes to M&A. It has done plenty of small bolt-on deals, including a handful last year. But efforts to play consolidator in the industry, which does everything from testing toys’ safety to certifying oil-drilling kit, hit a roadblock when French peer Bureau Veritas spurned it in favour of tying up with SGS of Switzerland — another deal that ultimately fell apart.

After their jump, Intertek’s shares are now trading some 3 per cent below EQT’s £51.50 per share offer, implying that investors don’t yet see much chance of it plonking down a whole lot more cash before its put-up-or-shut-up deadline in mid-May — or of another bidder being flushed out.

Still, the board has some levers to try to squeeze a little bit extra out of its private equity suitor. EQT’s offer — while generous beside last week’s share price — is less than 10 per cent higher than where shares were trading in early March, before disappointing growth rates pummelled them. It might not take much to produce a deal that’s harder for the company to refuse.