The AI trade just ran into one of its biggest reality checks yet.

More than $1.3 trillion in market value vanished from the world’s largest semiconductor companies in just a few trading sessions as investors pulled back from one of Wall Street’s hottest bets. Nvidia alone erased $238 billion in value since Friday’s market close. Memory chip leaders SK Hynix, Samsung Electronics, and Micron joined the selloff, signaling a sharp shift in sentiment across the industry that has fueled the AI boom.

The sell-off is not driven by a decline in demand for AI chips. It reflects growing questions about something far more fundamental: whether the trillions of dollars flowing into AI infrastructure will produce returns that justify the spending.

The decline spread across North America, Asia, and Europe, pulling down companies that sit at nearly every level of the AI hardware supply chain, from chip designers and manufacturers to memory suppliers and semiconductor equipment makers.

According to a CNBC analysis using FactSet data, the world’s 20 most valuable chip stocks have lost a combined $1.3 trillion in market capitalization since Friday’s market close.

Nvidia led the losses with a $238 billion decline. Taiwan Semiconductor Manufacturing Co. shed roughly $119 billion, AMD lost around $110 billion, Micron dropped approximately $113 billion, Samsung Electronics erased about $173 billion, and SK Hynix gave up nearly $176 billion.

The pullback comes after a remarkable run. The Philadelphia Semiconductor Index, or SOX, climbed 92% over the past year as investors poured money into companies expected to benefit from soaring demand for AI computing. Even after losing nearly 20% over the past month, the index remains far above where it traded a year ago.

Michael Field, chief equity strategist at Morningstar, said the latest decline appears to reflect investor psychology more than weakening business fundamentals.

“This decline appears to be driven largely by sentiment rather than fundamentals,” Field said. “Simply put, it’s a loss of confidence. We continue to see upside in many AI names, but these are growth stocks, and, as such, much of their value comes from cash flows expected far out into the future, which requires a lot of faith from investors.”

Why Wall Street Is Suddenly Questioning the Economics of AI Infrastructure

Investors spent much of the past two years rewarding nearly every announcement tied to AI chips, data centers, and cloud infrastructure. That narrative is beginning to change.

Charlie Dai, vice president and principal analyst at Forrester, said investors are reassessing whether near-term financial results can support unprecedented AI infrastructure spending.

“Investors are reassessing whether near-term revenues can justify unprecedented AI spending levels, while some also worry about growing competition in chips and AI infrastructure,” Dai said.

He added that the market reaction is “less about weakening AI demand and more about a repricing of expectations after an exceptionally strong rally.”

Those concerns have surfaced at the same time that major technology companies continue raising spending plans. Alphabet recently increased its 2026 capital expenditure forecast as it races to build additional AI infrastructure, reinforcing the industry’s commitment to long-term investment. The question now is no longer whether companies will spend. Investors want to know how quickly those investments will generate profits.

That distinction carries broad implications across the technology sector. Data center operators continue ordering AI accelerators, utilities are adding power capacity, construction firms are building new campuses, and cloud providers are signing long-term infrastructure agreements. Public markets are beginning to place greater weight on the financial returns those projects will produce over the next several years.

The Selloff Spread Across Global Chip Markets

The pressure quickly extended beyond U.S. markets.

South Korea’s technology sector absorbed some of the heaviest losses. SK Hynix closed down 9.61% after falling more than 15% during trading, even after reporting record quarterly revenue and profit. Samsung Electronics dropped more than 5%. LG Innotek lost 10.89%, and Seoul Semiconductor declined 8.89%.

Japanese semiconductor companies followed the same pattern. Memory maker Kioxia fell 13.85%. Tokyo Electron dropped 10.59%, and SoftBank Group, whose Arm investment has made it a popular AI proxy, lost 6.95%.

Taiwan Semiconductor Manufacturing Co., the world’s largest contract chip manufacturer, closed down 3.51%.

Europe offered few places to hide. ASML fell 1.77%, ASM International lost 3.28%, and BESI finished modestly higher after a volatile session.

Kieron Poon, investment director of Asian equities at Aberdeen Investments, attributed much of the weakness to deleveraging in South Korea and softer sentiment across global technology stocks.

“The recent market pullback has brought valuations to more attractive levels, creating opportunities for us to add exposure to high-quality businesses at more reasonable prices,” Poon wrote.

AI Demand Hasn’t Disappeared

The selloff arrives just one day after semiconductor stocks came under renewed pressure as investors questioned the scale, financing, and expected returns of massive AI infrastructure projects. The weakness spread across companies involved in AI processors, networking hardware, memory chips, semiconductor manufacturing equipment, and related technologies. China’s growing semiconductor industry has added another layer of uncertainty by raising expectations for stronger competition and greater manufacturing capacity.

Few analysts believe AI demand itself is fading.

Technology companies continue ordering advanced chips. Cloud providers are signing multiyear infrastructure contracts. Developers continue training larger AI models that require vast computing resources.

David Riedel, founder and president of Riedel Research Group, described the latest pullback as a healthy adjustment after an extraordinary rally.

Investors, he said, are simply “giving back a little bit of the froth that was in the AI market.”

He added that memory chipmakers “will be fine” but “just have to give back some of those sudden gains.”

The shift marks a new phase for the AI investment cycle. The debate has moved beyond whether artificial intelligence will reshape industries. Wall Street is now asking a different question: how long it will take for hundreds of billions of dollars in AI infrastructure spending to translate into sustainable earnings.

That answer could determine the next chapter for the semiconductor industry and the broader AI economy.