Key points
- Samsung Electronics and SK Hynix together are roughly 47% of the whole KOSPI, and briefly passed 51% of the KOSPI 200 in May. That is two companies, out of about 880 on the exchange, holding close to half the weight of the entire market.
- KOSPI allows stocks to move as much as 30% in a single day, more than double what most people are used to in the US market. When a stock hits that limit, it locks there and nobody can sell out until the next session.
- South Korea's regulator let 16 new leveraged ETFs launch on Samsung and SK Hynix in late May. Within weeks assets grew from about $3 billion to $9.1 billion, and by June 23 the whole thing triggered a market-wide circuit breaker.
- Retail investors here, called "ants," are 15 million people doing 60 to 70% of all trading volume. Margin-based investing jumped 72.5% in 2025 alone, nearly double the pace in the US. Some leverage products, called CFDs, historically let ants borrow up to 10 times their money.
I grew up in Korea, and it is personal here in a way I don't think people in the US always understand. So when people ask me why KOSPI moves so violently compared to other markets, I don't think of it as one reason. It is really three things stacked on top of each other.
Two companies, half the whole market
Samsung Electronics and SK Hynix now make up around 47% of KOSPI's total value, and for a stretch in May they were briefly above 51% of the KOSPI 200 index specifically. The exchange has about 880 listed companies. Two of them are basically half.
This is not normal, even compared to other concentrated markets. In the US, Apple, Microsoft, and Nvidia combined are large, but nowhere close to half of the S&P 500, which has 500 companies spreading the weight around. KOSPI's concentration exists because Korea's economy really is built this way. A small number of family-controlled conglomerates, called chaebols, generate most of the country's growth, and the memory chip business specifically happens to sit inside just these two companies. Samsung and SK Hynix are the ones making the HBM and DRAM chips that go into nearly every AI server. We explained how HBM chips work here if that part is new to you.
There's also something people call the "Korea discount," and I hear that phrase constantly. International money managers give Korean stocks lower valuations than similar companies elsewhere. Chaebol governance is genuinely hard to read from outside. Minority shareholders don't get treated the same as the controlling family, and most people here already know it. Dividend payouts tend to run lower too. So the rest of the market trades cheap while two chip giants carry the index. When those two stocks fall, the whole index falls with them, because there is nothing else on the exchange big enough to offset it.
How a 12% chip selloff becomes a 10% index crash
I want to show you one real week instead of just describing it, because the math only makes sense once you see it laid out.
| When | What happened |
|---|---|
| June 22, 2026 | SK Hynix passes Samsung Electronics in market value, the first time in over 25 years |
| June 22, 2026 | FSS Governor Lee Chan-jin says he regrets approving the new leveraged ETFs |
| June 23, 2026, 9:06 a.m. | Korea Exchange triggers a sell-side sidecar, a 5-minute pause on program sell orders, as the chip selloff from Wall Street hits Seoul |
| June 23, 2026, around 2:33 p.m. | KOSPI falls to about 8,378, down 8.1% on the day, triggering a Level 1 circuit breaker and a 20-minute trading halt |
| June 23, 2026, close | KOSPI closes at 8,203.84, down 9.99% on the day, one of its largest single-day drops on record |
| June 23, 2026 | SK Hynix falls 12.47% to 2.56 million won, its steepest drop since December; Samsung Electronics falls more than 10% |
| June 23, 2026 | Some of the new leveraged ETFs fall more than 25% |
Because Samsung and SK Hynix are close to half the index, an average drop of around 11% in the two of them mathematically drags KOSPI down roughly 5% before any other stock even joins in. The index actually closed down almost 10%, so the rest of the market sold off just as hard.
Korea also has a rule that makes this worse, one you don't really see in the US. Individual stocks can move up to 30% in either direction in a single day, a limit raised from 15% back in 2015. Once a stock hits that 30% floor, it locks there, and nobody can get out until the next session opens. A shorter pause called a Volatility Interruption kicks in around 10% single-stock moves too, and those pauses hit an all-time record of 29,357 times in just the first half of 2026.
Where the 10 times leverage comes in
Concentration and price limits explain why one bad day gets so extreme. Leverage is a separate problem. It is why bad days keep coming back.
There isn't just one leverage tool here. There are three, stacked on top of each other, and each one got worse this year.
| Tool | Leverage | What happened in 2026 |
|---|---|---|
| Single-stock leveraged ETFs (Samsung, SK Hynix) | 2x the stock's daily move | Launched late May. Assets went from about $3 billion to $9.1 billion within weeks. Some fell more than 25% on June 23. |
| CFDs (contracts for difference) | Historically up to 10x | Tightened after a 2023 case used CFDs to hide over $500 million in fake trading. Traders now must put up 40% of the position themselves. |
| Ordinary margin loans | Varies by brokerage | Total borrowing hit about 38 trillion won by early June, up from 27 trillion won at the end of 2025, roughly a 40% jump in about five months. Across all of 2025, margin-based investing grew 72.5% for the year, nearly double the 36.3% pace in the US over the same year. |
On one bad day in May alone, brokerages sent out 91.7 billion won worth of margin calls in a single session. Here is the part that turns a decline into a crash. When prices fall and a lot of margin calls hit at the same time, the brokerage sells the stock for you, whether you want it sold or not. That forced selling pushes the price down further, which triggers the next round of margin calls right behind it.
The "ants," and why Koreans trade this hard
Korean retail investors have a nickname: 개미, which means "ants." It's not an insult. It's just the normal word people use for them, on the news and in everyday conversation. There are around 15 million of these individual investors in a country of about 51 million people, and they alone are responsible for 60 to 70% of all the trading volume on the exchange. Big institutions dominate US trading far more than they do here. Korea really is a nation of small individual traders, checking their phones between meetings.
I think this comes from something real about how people here relate to money and risk. Real estate got so expensive that a lot of younger Koreans feel like a stock trade, even a leveraged one, is a more realistic path to catching up than saving slowly. Watching the market is close to a national pastime here, almost the same intensity people put into school here. Leverage products keep selling out within weeks of launch. Trading volume hit a record 106.2 trillion won in a single day this past May, almost 60% above the average from the months before it.
I don't think Korean investors are reckless. I think they are ordinary people using the tools that are actually sold to them here, and those tools go up to 10 times leverage on a CFD and 2 times leverage on a brand new ETF, inside a market where two stocks already make up half the index. Put those numbers together and the swings stop looking strange to me.
Why this matters if you're watching from the US
If you hold SK Hynix's US shares, a DRAM ETF, or anything tied to AI memory chips, KOSPI's overnight session already moved before your own market even opens. We covered how that overnight feedback loop works here, and what an actual KOSPI circuit-breaker day did to Micron and the rest of the memory sector. The structure explained above is why these moves, from a market most Americans check maybe once a year, can end up on your own portfolio the next morning.
Figures in this piece are sourced from Korea Exchange data, Financial Supervisory Service statements, and reporting on the May-June 2026 leveraged ETF episode and margin trading trends. This is general information about market structure, not investment advice.
