When stock markets break: South Korea’s market warning every investor should have seen

Recently, South Korea’s stock market did something that almost never happens. The KOSPI (South Korea’s equivalent of the S&P 500) plunged nearly 13% in a single session, triggering an emergency circuit breaker for the second consecutive day. At its lowest point during that session, the index had fallen 44% from the all-time high it had set just one month earlier.

Billions of dollars erased. In a month. From a market that had been one of the hottest on earth. You might be wondering why you should care about South Korean stocks. Here’s why…

The South Korea’s market setup and how it got so dangerous so fast

South Korea spent the past two years becoming one of the world’s biggest beneficiaries of the AI boom.
Companies tied to advanced memory chips – the kind that power AI data centres, humanoid robots and the next generation of computing infrastructure – became the most sought-after investments in the region. Samsung. SK Hynix. The suppliers and component makers orbiting them. As AI enthusiasm accelerated globally, South Korean technology stocks rose with it and then some.

Then a new product arrived that changed the risk profile of the entire market almost overnight.

Newly launched leveraged ETFs gave ordinary retail investors a simple, accessible way to amplify their bets on those same AI and semiconductor stocks. These products are designed to deliver two or three times the daily return of the underlying index. On the way up, they feel like genius.

You’re making double the market’s gain without having to do anything complicated.

But leverage isn’t a free lunch. It’s a loan. And loans have a feature that equity investments don’t – they can be called.

What happened last Wednesday

AI and semiconductor stocks had already been weakening for several weeks – both in South Korea and in the US. The correction was orderly.
Then the orderly correction became something else entirely.

As prices fell far enough, investors who had borrowed money to amplify their positions were forced to sell. Not because they wanted to, not because their fundamental view had changed, but because the mechanics of leverage required it. Their positions had lost enough value that margin calls were triggered. Leveraged ETFs mechanically rebalanced into falling markets. Forced selling hit an already declining market.

Those sales pushed prices lower. Lower prices triggered more forced selling. More forced selling pushed prices lower still. The self-reinforcing cycle that leverage always produces in a falling market accelerated into a single day of extraordinary destruction.

Notice what didn’t change last Wednesday. AI didn’t become less important. Demand for advanced memory chips didn’t disappear. The long-term outlook for Samsung and SK Hynix didn’t deteriorate overnight. The fundamentals were largely intact.

What changed was positioning. When too many investors crowd into the same trade using borrowed money, even a healthy correction can become far more severe than fundamentals alone would ever justify.

The market stopped pricing businesses and started pricing forced liquidation.

This has happened before

History doesn’t repeat exactly. But it rhymes with unsettling regularity.

In 1997, the collapse of Thailand’s currency looked, at first, like a local problem in a small market most global investors weren’t watching. Within months it had exposed financial weaknesses that spread across Asia and rippled into global markets – one of the most significant financial crises of the late 20th century.

In 2015, China’s domestic A-share market collapsed after a period of extraordinary retail investor leverage – margin lending had reached unprecedented levels during a two-year bull run. When the collapse came, forced selling created exactly the same self-reinforcing dynamic South Korea experienced last Wednesday. The Shanghai Composite fell 45% in three months.

In both cases, the early warning signs were in a market most Western investors weren’t watching closely. In both cases, the leverage (not the fundamentals) was what turned a correction into a crisis.

The stock market lesson that keeps needing to be relearned

There’s a critical difference between an ordinary market correction and one driven by forced selling.

Markets fluctuate every day as investors react to earnings, economic data, interest-rate expectations and geopolitical events. That’s normal.

That’s healthy. Those fluctuations are, in fact, the mechanism by which markets price information correctly over time.

Leverage changes the equation completely. Investors using borrowed money don’t always have the luxury of waiting for markets to recover. Once prices fall far enough, they’re forced to reduce positions regardless of what they believe those investments are actually worth. Selling becomes disconnected from fundamentals and begins feeding on itself. The market is no longer pricing businesses. It’s pricing fear and margin calls.

That’s what South Korea showed us last Wednesday. And the reason it matters – even for investors who own no Korean stocks and have never heard of the KOSPI – is that the same dynamic can emerge wherever leverage has accumulated in crowded trades.

Right now, in 2026, AI and semiconductor names are among the most crowded, most leveraged trades in global markets. Not just in South Korea. In the US. In Europe. In retail portfolios and institutional accounts simultaneously. The products that allow ordinary investors to take leveraged bets on AI have proliferated as quickly as the AI theme itself.

South Korea showed us what happens when those leveraged positions begin to unwind simultaneously.
Simply put – the leverage that created South Korea’s collapse didn’t build up overnight. It accumulated gradually, while the trade was working – until it wasn’t.

The most dangerous trades are never the ones that look dangerous at the time. They’re the ones that look unstoppable right up until the moment the selling starts, and the borrowed money has to go home.

Before you add to any position right now (in AI stocks, in semiconductor names, in anything that has become the consensus trade of the moment), ask yourself one question.

How much of the current price reflects fundamentals, and how much reflects borrowed money that still needs to unwind?

The answer to that question is more important than any earnings forecast or price target you’ll read this week.