AI’s Harder Market Test

A Global AI Trade Runs Into a Harder Test

The market’s faith in artificial intelligence has not vanished. But in recent days, investors across Asia and the United States have shown that faith alone is no longer enough.

A fresh sell-off in semiconductor and AI-linked stocks swept through global markets in late July, dragging down some of the sector’s biggest winners even after strong earnings and rapid growth. In South Korea, shares of SK Hynix tumbled after the memory-chip maker posted record quarterly results that would, in another market, have been a triumph. In Japan, SoftBank slid sharply as investors dumped AI-related holdings across the region. In the United States, losses in chip stocks including AMD, Intel and Micron extended a rout that had already gathered force on Wall Street.

The common message from investors was strikingly unforgiving: strong results were no longer sufficient if they failed to exceed already extraordinary expectations.

That shift is significant because AI-related stocks have been among the defining market trades of the past several years, lifting indexes, driving retail speculation and tying together markets from Seoul to Silicon Valley more tightly than before. Now, as valuations come under scrutiny, those same linkages are helping spread the pain.

Korea Becomes a Pressure Point

South Korea has emerged as one of the clearest examples of how concentrated and leveraged the AI trade had become.

SK Hynix, a central supplier of high-bandwidth memory chips used in AI systems, reported surging profit and revenue, underscoring how demand for AI infrastructure remains robust. Yet its shares fell anyway, suggesting investors had already priced in even more aggressive gains. Samsung Electronics, another pillar of the country’s technology complex, also declined.

The market fallout was especially severe for retail investors who had piled into leveraged exchange-traded funds tied to single stocks, particularly Samsung and SK Hynix. Those products magnified losses as the chip sell-off deepened, leaving many individual investors nursing heavy damage. The episode became serious enough that South Korea’s finance minister publicly apologized as the losses mounted.

The government had already begun trying to cool the market. On July 16, regulators tightened rules on single-stock leveraged products, raising the minimum cash deposit for retail investors to 30 million won, from 10 million won, pausing new product listings and curbing the marketing of existing funds. But the measures came after leverage had already built up significantly.

Officials had warned that Samsung and SK Hynix together accounted for roughly 52 percent of the KOSPI’s market capitalization by mid-July, a degree of concentration that made the broader market unusually vulnerable to any reversal in AI enthusiasm. Retail investors were also still carrying tens of trillions of won in margin exposure, making South Korea an amplifier of a trade that had started as a global bet on AI demand.

The Contagion Spreads Across Markets

What might once have looked like a local market correction now appears more global.

Data from Rayliant cited in recent market analysis showed the 60-day correlation between South Korea’s KOSPI and the Nasdaq 100 had risen to about 0.50, the highest level since 2021. That is a notable sign that Korean and American technology stocks are moving more in tandem, reflecting the way AI optimism has come to dominate both markets.

The regional weakness was evident beyond Seoul. In Tokyo, SoftBank fell sharply as AI-linked investments came under pressure. Across the United States, chip stocks remained under strain, with AMD, Intel and Micron extending losses as investors reassessed how much future growth was already embedded in valuations.

Not every corner of the technology sector moved lower. Some Chinese internet companies listed in Hong Kong traded higher, bucking the broader weakness. But that divergence only highlighted the central theme of the sell-off: the sharpest pain was concentrated in the companies most directly tied to the AI build-out and in the stocks where expectations had become most stretched.

When Strong Earnings Are Not Enough

The latest market reaction reflects a deeper turn in investor psychology.

For much of the AI boom, companies exposed to data centers, advanced memory, chip design and AI infrastructure were rewarded for showing almost any evidence of participation in the trend. More recently, however, the market’s demands have escalated. Investors are no longer asking simply whether revenue and profits are growing. They are asking whether they are growing fast enough to justify valuations built on the assumption of sustained, near-flawless expansion.

That is what made the reaction to SK Hynix so telling. Record earnings and explosive growth would ordinarily reinforce confidence in the AI spending cycle. Instead, they prompted disappointment, a sign that the market may be transitioning from euphoria to a more exacting phase.

In the United States, a similar skepticism has surrounded chipmakers whose fortunes are tied, directly or indirectly, to data-center demand and AI infrastructure spending. The sell-off suggests investors are beginning to question whether the next leg of growth can keep matching the pace implied by stock prices.

The Bigger Questions for the AI Boom

The immediate trigger for the declines may be earnings that failed to clear a very high bar. But the deeper debate now revolves around the durability of the AI investment cycle itself.

Investors are trying to determine whether this is primarily a valuation reset after an overcrowded rally or the start of a broader reassessment of AI economics. Among the central questions: Can the world’s largest cloud companies keep increasing capital spending at the pace bulls expect? Will financing concerns around the enormous cost of AI infrastructure begin to weigh more heavily on the sector? And could rising competition, including from Chinese memory-chip producers, put pressure on pricing and margins?

In South Korea, another question hangs over the market: whether tighter rules on leveraged retail trading can prevent further forced selling. If prices continue to fall, leverage itself could become a driver of additional volatility, especially in a market where so much weight is concentrated in just a handful of chip names.

For now, the pullback does not amount to a verdict against artificial intelligence as a business theme. Demand for the hardware that powers AI remains strong, and the industry’s biggest players are still posting formidable growth. But the market is making clear that a new phase has begun — one in which even exceptional results may disappoint if they do not surpass expectations that have become almost impossibly high.