AI-Chip Rout Pulls Nasdaq Down Nearly 4%

The most reliable trading signal from this session is not the Nasdaq's exact percentage move - which sources report inconsistently depending on what point in the session they measure - but the KOSPI's 10% crash and the more-than-12% declines in Samsung Electronics and SK Hynix, since those two companies supply the large majority of the high-bandwidth memory chips used in AI data centers and are a more direct read on AI-infrastructure demand sentiment than a broad US index close.

A rout in AI and semiconductor stocks that began in Asia spread to US mega-cap tech on June 23, 2026. Independent market-data reporting (Yahoo Finance, MLQ, CNBC) confirms the Nasdaq Composite closed down 2.2% (about 579 points) at 25,587, while the S&P 500 fell 1.4%. South Korea's KOSPI closed down roughly 10% at 8,203.84, triggering a market-wide, 20-minute circuit-breaker; Samsung Electronics and SK Hynix, which together account for about 48% of the KOSPI's market value, both fell more than 12% as foreign investors sold a net 5.79 trillion won (about $3.8 billion) of Korean shares. In the US, Micron Technology - reporting quarterly earnings the next day - led chip-stock losses with a 13.2% drop; AMD fell 5.8%, Intel 6.1%, and Nvidia about 4.1% to roughly $200, dropping below its 50-day moving average. The VanEck Semiconductor ETF fell 7.1%, its steepest single-day decline in months.

This story originally cited InvestorIdeas for a deeper Nasdaq decline (3.94% to 25,773.73) and a sharper VIX move (+33% to 20.44); those figures do not match the session-close numbers corroborated by Yahoo Finance, MLQ, and CNBC, and most likely describe an earlier, more volatile point in the session (intraday or afternoon trading) rather than the final close. The KOSPI, Samsung, and SK Hynix figures are consistent across all sources reviewed.

Markets commonly treat memory-chip makers as a leading indicator for AI hardware supply and demand, because large memory purchases feed training and inference capacity; companies and ETFs exposed to memory and AI compute tend to show higher beta in rapid risk-off episodes. Circuit-breakers in concentrated, tech-heavy indices can accelerate cross-market contagion when a handful of large-cap names account for a large share of index capitalization, as Samsung and SK Hynix do on the KOSPI.

The divergence between mega-cap names with proven AI revenue (Microsoft rose 1.8% the same session) and hardware-heavy names bearing the brunt of the selloff suggests investors are starting to differentiate within the AI trade rather than de-risking uniformly. For practitioners, rapid re-pricing in memory and GPU-related equities can translate into shorter-term funding and hiring headwinds for hardware-heavy projects and tighter risk premia on infrastructure spending.

Micron's earnings the following day as the first major data point on AI memory demand since the selloff; whether South Korean retail investors' contrarian buying (a reported net 11.11 trillion won during the session) holds up against continued institutional selling; and whether circuit-breaker frequency and index concentration in a handful of AI-hardware names becomes a recurring feature of 2026 volatility episodes.