Wall Street Shifts Focus from AI Giants to Infrastructure Companies

As major artificial intelligence companies see their stock prices decline, investors are turning their attention to infrastructure firms that support AI development. New investment funds are targeting companies that build data centers, provide power, and manufacture chips needed for AI operations.

Wall Street’s enthusiasm for major artificial intelligence companies is cooling off, prompting investors to redirect their focus toward infrastructure businesses expected to profit from AI-related spending, according to a new Reuters analysis.

Following massive growth in previous years, technology giants like Alphabet and Amazon have experienced significant stock price drops as market participants question whether returns from their enormous AI investments will support current high valuations. Investment managers report that to capitalize on this spending boom, investors are now targeting the businesses receiving those investment dollars — semiconductor manufacturers, data center construction companies, and utility providers supplying the essential infrastructure powering the AI transformation.

Several infrastructure-related stocks, including Caterpillar, optical communications firm Lumentum, and data storage company Western Digital, have achieved double-digit percentage increases this year. Meanwhile, the S&P 500 has delivered just 0.52% returns, and the Roundhill Magnificent 7 ETF, tracking major AI companies, has dropped 7.3%.

This market performance is driving exchange-traded fund companies including BlackRock, VistaShares, and Impax Asset Management to redesign their investment products and introduce new offerings, with some focusing on diverse and increasingly specialized AI infrastructure investments.

Adam Patti, CEO of VistaShares, explained their strategy for the Artificial Intelligence Supercycle ETF launched in December 2024: “Our goal is that every time someone like Meta or Amazon invests in a data center, the cash registers ring across our portfolio.” The fund gained 58.4% in 2025 and has risen 16.87% this year.

Though the VistaShares ETF holds AI leader Nvidia, the semiconductor company’s portfolio weight is less than half of South Korea’s SK Hynix, whose processors power data centers. Other major holdings include chip manufacturers Micron and Intel.

“When Meta says that it’s going to spend $100 billion, it’s going into these companies,” Patti noted.

Similarly, BlackRock’s iShares A.I. Innovation and Tech Active ETF now allocates 74% of its $8.8 billion in assets to AI infrastructure investments, spanning from chip companies training AI systems to power providers, increasing from 59% twelve months ago. Jay Jacobs, BlackRock’s U.S. head of equity ETFs, said this shift reflects “where the revenues are right now.”

Strong performance from holdings such as Fabrinet and Monolithic Power Systems has driven the fund’s returns to 3.2% this year. The BlackRock fund has attracted $7.9 billion in fresh investment over the past year, based on VettaFi data.

February alone saw two infrastructure ETF launches. Impax Asset Management transformed one of its mutual funds into the Impax Global Infrastructure ETF, while alternative investment manager Harrison Street Asset Management introduced an AI-focused ETF emphasizing electrification.

Robert Becker, chief investment strategist at Harrison Street, emphasized the power challenge: “Securing reliable power sources is one of the biggest constraints in moving forward with all the AI data centers needed.”

Ed Farrington, Impax’s president of North America operations, described infrastructure as a method to diversify overall stock portfolios and what has been a highly concentrated investment strategy for years.

While the major technology companies have maintained strong revenue growth, investors note this success primarily stems from their established business operations, which finance AI capital expenditures expected to reach approximately $630 billion this year.

The search for undervalued infrastructure companies positioned to benefit is directing some investors toward specialized market segments.

Ari Sass, president and portfolio manager of M.D. Sass Investor Services, said companies he previously considered “stealth” AI investments are gaining attention, particularly those helping supply the massive energy requirements for semiconductor manufacturing facilities and data centers.

Quanta Services, providing construction and maintenance for electric utilities, has climbed 24.17% year-to-date.

The Tortoise AI Infrastructure ETF, launched in October, invests in companies like century-old Wisconsin-based Modine Manufacturing, which began by producing radiators for agricultural equipment and has transitioned to supplying data center cooling systems. Its stock has increased 19.25% this year.

As more investors enter the AI infrastructure market, some experts are issuing cautionary warnings. They reference the fiber optic network companies that failed after excessive investment to support internet companies in the 1990s as a historical warning.

Michael Reynolds, vice president of investment strategy at Glenmede, offered this perspective: “It looks as if the spending on AI buildout is coming from financially stronger companies, but at the same time, valuations for anything with AI exposure are getting a bit rich. Everyone needs to exercise some caution.”