This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.
Korea’s KOSPI roughly doubled in 2026, closing above 9,000 for the first time in mid-June — up about 95% on the year, the best-performing major market on the planet. (Verify the exact close and return before relying on them.) So here’s the question everyone’s really asking: is this Korea’s long-awaited AI moment, or is it 1999 all over again? The honest answer is “both, in different layers” — and if you can’t tell the engine from the fuel, you’ll end up buying the wrong one.
The short answer: separate the engine from the fuel
Two very different things pushed this index to records, and they don’t share a fate. One is an earnings engine — the AI-memory super-cycle — and it’s real and durable. The other is fuel: record retail leverage and an index propped up by just two stocks. That’s psychology, and psychology changes direction. “Should I buy now?” depends entirely on what you’re buying. Owning earnings-backed AI-chip exposure for the long run is a completely different game from chasing record highs on borrowed money — and the latter should remember how 1999 ended. This is not financial advice.
What actually happened
The KOSPI started 2026 in the 4,000s and hit roughly 9,064 by June 19 — about +95% on the year. The direct trigger is AI: SK Hynix announced it had shipped next-generation HBM4E 12-layer memory samples to key customers and jumped sharply, dragging Samsung Electronics and the rest of the chip chain up with it. But the same month also saw intraday drops of 7-8%. (Verify.) This isn’t a calm uptrend — it’s the textbook look of a parabolic market, where big gains and big air pockets arrive together.
🎩 Under the Gat — If you’re wondering how the index can hit records while foreigners dump stock by the trillion, the answer is one Korean word: bittu (빚투) — “investing with borrowed money.” Domestic retail is absorbing what foreigners sell, and a lot of it is on margin. The question isn’t whether the music stops. It’s who’s holding the position when it does.
Both sides, fairly — bull vs. bear
The bull case (the engine is real). AI memory isn’t a passing theme; it’s a structural super-cycle, and SK Hynix’s HBM throws off actual profits. Unlike the dot-com darlings, Korea’s index leaders make money. Add Korea’s Value-up corporate-reform push and hopes for an MSCI developed-market upgrade, and bulls argue the re-rating is justified.
The bear case (the fuel is dangerous). The worry isn’t earnings — it’s how the price got here. First, retail margin debt is at a record: the credit-loan balance hit roughly ₩36.5 trillion (~$26.9B) by mid-June, about double a year earlier. (Verify with KOFIA.) Second, the index is dangerously concentrated: as of June 19, Samsung Electronics (~28%) and SK Hynix (~27%) together made up roughly 55% of the entire KOSPI’s market cap — if those two wobble, the whole index wobbles. Third, foreigners have been net sellers of a record ~₩103 trillion in 2026 (more than in 2008 or 2020), which bears read as smart money heading for the exits. (Verify.)
🎩 Under the Gat — Don’t confuse the engine with the fuel. HBM earnings are the engine. Bittu is the fuel. An engine moves the car; leaking fuel starts a fire. There’s a phrase going around Korean retail circles right now — roughly, “I’d rather risk total collapse than miss the rally.” Admirable nerve. Margin calls don’t grade on nerve.
The anchor for US readers: this rhymes with the Nasdaq in 1999
The closest mental model for an American investor is the Nasdaq in 1999 — a genuine technological boom (then the internet, now AI memory) wrapped around a leverage-fueled mania. On the retail-margin angle alone, it also rhymes with the 2020-21 Robinhood era, when small investors chased highs on borrowed money.
| KOSPI 2026 | Nasdaq 1999 (US anchor) | |
|---|---|---|
| What’s driving it | AI memory / semiconductors (HBM) | The internet / dot-coms |
| Earnings support | Leaders make real profits (Hynix HBM) | Many had no profits |
| Retail leverage | Record margin debt (~₩36.5T) | Surging margin loans |
| Concentration | Samsung + Hynix ≈ 55% of the index | Top tech heavy, but more spread out |
| Flows | Foreigners selling ↔ retail buying on margin | Retail pouring in late |
| Decisive difference | Two-stock concentration + real earnings | Many no-profit names + broader spread |
Not a perfect parallel — in 1999 a lot of companies had no earnings, while Korea’s index leaders today genuinely do. The KOSPI’s real weak spot isn’t no-profit hype; it’s the combination of two-stock concentration and retail leverage. So the sharper question isn’t “is it a bubble?” — it’s “what gets dragged down if those two stocks fall?”
The Korean texture: the ant army and “bittu”
To understand this rally you have to meet the donghak gaemi (동학개미) — literally the “righteous ant army,” Korea’s legion of small retail investors. The name was born in the 2020 crash, when individuals soaked up the Samsung shares foreigners were dumping. The 2026 version is more aggressive: bittu (빚투), “investing with borrowed money.” As institutions and foreigners take profits and leave, retail is using credit to chase the highs. There’s a generational backdrop here — with the property ladder feeling out of reach, many younger Koreans treat the stock market as the last available route to moving up.
🎩 Under the Gat — Outsiders see numbers. Locals read the mood. Bittu isn’t a data point; it’s an emotion — and positions bought on emotion are the first to be sold when the mood cools. Want AI exposure? Good. Just don’t own it on margin.
Risks worth naming (not predictions — things to watch)
- A leverage unwind. Record margin debt means a drop can trigger margin calls, which force selling, which trigger more margin calls. (Verify balances with KOFIA.)
- Two-stock dependence. With Samsung and SK Hynix at roughly 55% of the index, the whole market is hostage to HBM pricing and the memory-chip rivalry (Micron, Samsung’s catch-up).
- Foreign flows and FX. Persistent foreign selling plus a moving won/dollar rate swing returns further for foreign-based investors.
- Stretched corners. Battery and materials names trading far above their earnings carry especially violent volatility. (Verify multiples.)
The bottom line: the boom is real, and so is the leverage. They are not the same thing — and telling them apart is the whole job here.
— Mr. Gat 🐂
This is not financial advice. Past performance does not guarantee future results, and leveraged (margin) investing can cause losses exceeding your principal.
Frequently Asked Questions
The KOSPI is at a record high — is it safe to buy now?
It depends entirely on what you buy and how. Taking long, unleveraged, diversified exposure to earnings-backed AI semiconductors is a very different risk than chasing the index at record highs on margin. The latter is extremely volatile. This is not financial advice.
If foreigners are selling, why does the index keep rising?
Domestic retail investors (the donghak gaemi) plus ETF and pension flows are absorbing the foreign selling. Foreigners net sold a record ~103 trillion won on the KOSPI in 2026, yet the index kept climbing. A large and growing share of that retail buying is on margin.
Is this a bubble?
It’s not a simple yes or no. The earnings engine is real — unlike the dot-com era, Korea’s index leaders actually make money. But record retail leverage and extreme concentration in two stocks are genuine vulnerabilities. The real question is “what breaks if those two stocks wobble?”
How can a foreigner get exposure to this rally?
Through a US-listed ETF like EWY (iShares MSCI South Korea) or by buying Korean shares directly via a broker such as Interactive Brokers. Just know that a “Korea ETF” is now close to a concentrated bet on memory chips, because Samsung and SK Hynix dominate the index.
What signals tell me the rally is overheating or cooling?
Watch the retail margin-loan balance, whether foreigners turn net buyers, Samsung’s and SK Hynix’s HBM pricing and margins, and intraday volatility (how often the index swings hard). Watch these internal gauges, not the headlines.
This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.