Why the Nikkei 225 and KOSPI Crashed on July 28 — and What It Means for AI Stocks

Nikkei performance. The Nikkei 225 tumbled more than 2,500 points intraday, with AI-linked names like Kioxia, Tokyo Electron, and Advantest leading the losses .

Broader Structural Concerns Driving the Ongoing Selloff

Circular Financing and AI Infrastructure Risks

Investors are increasingly worried that Nvidia's top customers — the hyperscalers — are funding massive AI capex through debt and could face financing strains if ROI disappoints, potentially creating a "circular" cycle where chip demand depends on borrowed money . A Reuters report cites "mounting concerns over financing risks tied to AI infrastructure" as a key factor in the July 28 selloff .

The unwind of a leveraged South Korean semiconductor ETF on June 23 highlighted the dangers of excessive retail leverage and crowded positioning in AI trades . Analysts have described the move as the beginning of a new phase of the AI trade — one defined by harder questions about valuations rather than blind confidence in the AI buildout narrative .

Stretched AI Valuations Despite Strong Earnings

Samsung reported a 19-fold jump in Q2 operating profit to 89.4 trillion won ($58.4 billion), yet its shares fell nearly 7% on that news alone . Investors concluded that future AI-driven growth was already priced in . TSMC's strong results likewise "failed to clear investors' lofty expectations," triggering a fresh wave of selling on July 16 .

The KOSPI entered bear territory in early July after being the world's best-performing stock market this year, underscoring how quickly sentiment turned on AI concentration risk . Samsung Electronics and SK Hynix together account for more than half of the KOSPI's market capitalization, making the index acutely vulnerable to AI sentiment swings .

Upcoming Big Tech Earnings as Volatility Catalysts

Every major earnings release has become a volatility event. Meta's plan to sell AI computing capacity raised overcapacity fears that sank Asian chips on July 2 . TSMC's earnings on July 16 failed to reassure, as investors focused on rising costs and fatigue over the years-long AI boom .

Markets are bracing for further inflection points from Nvidia, Apple, and other hyperscalers, with each report testing whether AI revenue growth can justify the industry's trillion-dollar-plus capex pipeline . Goldman Sachs' Prime Services desk reported that hedge funds have been reducing tech exposure ahead of earnings .

Key Takeaways

The July 28 sell-off was not a one-day panic but the latest and most violent episode in a longer repricing of AI and semiconductor stocks that has been building since early June 2026. The simultaneous impact of a new Chinese competitor, stretched valuations, and circular financing fears suggests the correction may have further to run, especially if upcoming Big Tech earnings fail to deliver a decisive positive surprise.