Global tech stocks plummet: Another stress test for the AI bull market

```June 23, Seoul, South Korea.

At 2:00 PM, the Korea Exchange triggered a circuit breaker. After the KOSPI index plunged 8%, trading was halted for 20 minutes. When trading resumed, the index continued to fall and ultimately closed down 9.99% at 8203.84 points.

Several numbers outline the intensity of this selloff——

Samsung Electronics dropped 12.31%, SK Hynix fell 12.47%. These two companies alone accounted for about 71% of the KOSPI’s decline for the day. The Nikkei 225 fell 3.55%, losing the 70,000 level. Nasdaq 100 futures fell 3.01%, the Philadelphia Semiconductor Index closed down 7.7%. TSMC dropped over 5% pre-market, Micron more than 8%, AMD, Intel, Applied Materials, ARM, and ASML all dropped more than 7%.

The collapse of leveraged ETFs was even more eye-catching: Korea 3x long ETF plummeted 32% in a single day, 3x long semiconductor ETF slumped 17%.

KOSPI’s single-day drop ranks in the top five in Korean market history. The last comparable crash was in October 2008.

But 2008 was amid an obvious Great Recession. In 2026, the global economy is growing, an AI revolution is in full swing, and the KOSPI was still one of the world’s top performing major indices at the start of the year—until this crash.

So the real question is: What happened, and why?

Trigger: Resonance of Three Signals

Looking back at the timeline, the direct trigger for the crash was the resonance of three signals within 24 hours.

First: SK Hynix slows HBM4 capacity expansion.

On the morning of June 23, Korean media reported that SK Hynix was slowing its HBM4 capacity expansion, shifting focus to general DRAM. HBM (High Bandwidth Memory) is essential for AI chips, with SK Hynix and Samsung the only global mass suppliers. Market consensus was nearly unanimous that HBM4 would be in “short supply.”

HBM4 is one of the clearest bottlenecks in the global AI infrastructure race. Whenever markets begin to doubt the severity of this bottleneck, the result is usually a repricing of faith.

Second: Profit taking ahead of Micron’s earnings.

Micron will release quarterly earnings on Wednesday (June 25). Boosted by its full-stack partnership with Anthropic, Micron’s stock hit an all-time high, up over 300% this year. Goldman Sachs’ trading desk noted, “Investor expectations are extremely high, which creates conditions for taking profit ahead of the report.”

In an expectation-driven market, position adjustments before earnings often have more impact than the report itself.

Third: Korean regulators warn about leveraged ETFs.

On June 22 (the day before the crash), the head of Korea’s Financial Supervisory Service, Lee Chan-jin, publicly stated he “regretted” not stopping the issuance of single-stock leveraged ETFs tracking Samsung and SK Hynix, saying their function was “almost nothing but allowing brokers to profit at retail investors’ expense.”

The timing was brutally precise. When regulators publicly acknowledge structural problems, it triggers panic selling.

Amplifiers: The “Three Levers” of the Korean Market

The destructive force of these three signals was so great because they hit a market already deeply bound by leverage.

This round of the Korean AI bull market had three amplifiers working at once.

Amplifier One: Retail margin hit record highs.

Korean retail investors are known for their risk appetite. This round saw leverage reach unprecedented heights. Retail margin balances climbed to record highs before the crash. A post-crash Goldman Sachs report bluntly stated: “Korea’s market gains are increasingly dependent on retail investors as marginal buyers.”

In a leveraged cycle where rising markets attract more buyers, once marginal buyers turn around, a reverse cycle of falling prices and more selling is triggered.

Amplifier Two: Single-stock leveraged ETFs ballooned to $30B.

This is a unique issue in Korea. 16 domestic single-stock leveraged ETFs have $9.1B in assets; CSOP’s 2x long ETFs for SK Hynix and Samsung on HKEX total around $21B—combined, over $30B. Of the domestic products, 92% of holdings are from retail.

Single-stock leveraged ETFs have a fatal structural feature: they require daily rebalancing. When the underlying stock falls, the ETF must sell even more stock to maintain leverage, forming a self-reinforcing sell-off. When regulators hint at possible curbs, these products—and their underlying holdings—are the first to be sold.

The FSS estimates these products generated $3B–$6.4B in transaction commissions. Considered measures include: higher barriers for retail, tougher education tests, caps on ETF size, new product limits, and stronger suspensions when ETF prices deviate from NAV.

Whether these measures take effect or not, the signal is clear: regulators believe a significant part of the current rally is detached from fundamentals.

Amplifier Three: National Pension Service became an unexpected seller.

The National Pension Service (NPS)—Korea’s largest pension fund—sold a net $1B KOSPI stock in the six days before the crash, with net sales of $1.5B so far in June, the largest monthly sell since April 2021.

The NPS was merely rebalancing: sustained KOSPI gains had pushed domestic stock allocations over 30%, breaking the roughly 28.8% cap.

Crucially, in a market highly dependent on retail capital, when NPS—once the key stabilizing buyer—became a seller, nobody was left to "catch" the selloff.

Goldman analysts put it: “The so-called theoretical constraints have become observable liquidity realities.”

Bubble Debate: When Will It Burst, How Deep Will It Fall

After the crash, a debate about an “AI bubble” naturally emerged.

Chris Cha, head of high-touch trading at Goldman Korea, gave a clear qualitative comment on the day of the crash—liquidity exhaustion: “I remain constructive on the memory cycle and think KOSPI is undervalued. But recent gains are increasingly driven by technically minded buyers, making it much more vulnerable to momentum interruptions.”

In other words: the medium- and long-term logic hasn’t changed, but the short-term market structure is so fragile it’s at a tipping point.

BofA’s quant signals: Nasdaq approaching bubble territory.

Days before the crash, BofA strategists noted their bubble risk indicator showed Nasdaq 100 near 0.8, “usually signaling elevated near-term tail risk.” Tech and chip stocks “showed extreme bubble-like price action.”

But BofA also noted: “The AI bubble may take years to fully form. Historically, this indicator signals interim corrections, not trend reversals.”

Li Bei’s warning: The trigger conditions have appeared.

Private fund manager Li Bei also warned investors about the AI sector. In her view, “the trigger conditions for the AI bubble to burst have already appeared.”

Three voices, three timeframes: Li Bei says “now,” Goldman says “don’t panic,” BofA says “go higher—but a drop first.”

Notably: In a market deeply bound by leveraged ETFs, retail margin, and momentum trading, a “correction” and a “bubble burst” can look similar in price action. A 10% daily drop already triggered the circuit breaker. If there’s another 10%, what happens?

Mirror: SpaceX Tells the Same Story

If you shift your focus from Seoul to New York, you see a mirror image.

SpaceX, after going public, crashed for three consecutive days—over 16% on June 19, about 5% more on June 22, and continued falling on June 23—about $600 billion in market value wiped out in three days, falling below its first-day opening price of $150 to about $147.

Even more intriguing: As the stock plummeted, SpaceX announced its first bond issuance—raising $20 billion for AI infrastructure. Typically, companies borrow when in financial hardship. SpaceX is the opposite: it needs continuous “ammunition” to fund its AI capex story, but falling stock prices have closed the equity-refinancing window—debt is its last option to keep the story intact.

The echo between SpaceX and Korean semiconductors reveals a common issue: The AI theme’s capital market narrative is shifting from “infinite imagination” to “calculating returns.”

When investors start crunching numbers—How much has HBM4 expansion slowed? How much have GPU rental prices fallen? When will AI revenue cover capex?—the market’s pricing logic changes.

Looking Ahead: Micron’s Judgment Day

After the crash, everyone is focused on one date: June 25.

Micron’s earnings—the most direct “judgment” for tech stocks after this plunge.

Bloomberg quotes Pepperstone strategist Dilin Wu: “Micron’s earnings this week are the real test. A strong result will be a direct positive for Samsung and SK Hynix—the numbers will show whether AI hardware trades are still sound.”

There are two logical scenarios:

Scenario One: Micron beats expectations. If Micron releases solid earnings and gives optimistic guidance, the current selloff could quickly reverse—as uncertainty is resolved, positions will be refilled.

Scenario Two: Micron disappoints. If Micron’s guidance is weak, the selloff will get fundamental validation—the liquidity problem will be redefined as a crisis of belief.

In scenario one, the KOSPI circuit breaker will be branded as a “technical stampede.” In scenario two, it will be “the turning point for the AI bull market.”

The answer Wednesday night will decide Asia’s opening direction on Thursday morning.

Conclusion

The June 23 crash can be technically explained as: a singular SK Hynix HBM4 news + structural fragility of leveraged ETFs + NPS’s surprise selling for rebalancing + risk-off moves ahead of Micron earnings—four factors resonating within 24 hours.

But that’s just the technical explanation.

The deeper issue is: When the AI bull market reaches this point—valuations at historic extremes, the driving force shifting from institutions to retail, trading structures deeply bound to leverage, with inflation and rate hike expectations raising the risk-free rate—a structural break is almost inevitable.

The KOSPI circuit breaker is a mirror.

The mirror shows: When all market participants are leveraged on the same narrative, the reversal of this narrative—no matter how fast or deep—should not be a surprise.

For investors, after Micron’s earnings, only one question remains: How much drawdown are you willing to accept to keep holding?

Risk Warning and DisclaimerThe market has risks. Please invest cautiously. This article does not constitute personal investment advice, nor does it consider the specific investment objectives, financial situation, or needs of any individual user. Users should consider whether any opinions, views, or conclusions herein fit their specific circumstances. You bear all responsibility for acting on this information. ```