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.
This is not financial advice. One person’s view — not a recommendation to buy or sell anything.
Straight answer first: Monday’s 8.95% crash in Seoul is not, on the evidence available today, the AI trade breaking. Korea’s benchmark index, the KOSPI, fell 8.95% and closed below 7,000 for the first time in about two months. Trading was frozen mid-afternoon by a circuit breaker — Korea’s market-wide halt, which trips when the index drops 8% — the seventh such halt this year.
And yet, on the same day: the Korean won barely moved (0.13%). Refiners and defense stocks rose, some by double digits. Korean retail investors bought ₩3.88 trillion (~$2.58B) of what foreigners and institutions were selling. That is not the fingerprint of a market discovering that AI demand is over. It is the fingerprint of a crowded, leveraged position coming apart in two stocks that happen to be more than half the index. Same chart, different order book.
Here’s what a US reader has to understand about this market: the KOSPI is less an index than a thin wrapper around two chip stocks. Samsung Electronics and SK hynix — the world’s two biggest memory-chip makers, and the main suppliers of the HBM (high-bandwidth memory) that sits next to every AI accelerator — together were 54.76% of KOSPI market cap as of June 19, and higher after July’s rally. Imagine an S&P 500 where two names are half the index. When those two fall 10.70% and 15.37%, the index has already lost roughly seven percentage points before you look at the other 900-odd names. Some of them, on Monday, went up.
The conventional wisdom: “Middle East war + AI peak-out = risk-off has arrived”
The headline logic was clean. US forces struck Iran for the fourth time in a week, transits through the Strait of Hormuz have collapsed to a handful of vessels a day against roughly 130 before the war, and Brent September futures jumped more than 4% to $78.82 as of 08:00 GMT, the highest since June 22. Layer a semiconductor cycle-top note on top, and a 9% drop in Asia’s flagship chip market reads like the moment global risk-off landed in Korea.
It’s a tidy story. The problem is that the prices actually printed in Seoul on Monday only support half of it.
What the data says
| What moved | Monday’s number | What the “risk-off” story predicts |
|---|---|---|
| KOSPI | 6,806.93 (-8.95%), below 7,000; 7th circuit breaker of 2026 | ✅ Consistent |
| Samsung Electronics (KRX: 005930) | ₩254,500 (~$169) (-10.70%) | ✅ Consistent |
| SK hynix (KRX: 000660) | ₩1,845,000 (~$1,227) (-15.37%) — worst day in 17 years | ✅ Consistent |
| USD/KRW | 1,503.4 (+2.0 won, +0.13%) | ❌ A true risk-off breaks the currency first |
| Refiners | SK Innovation (KRX: 096770) +10.69%, S-Oil (KRX: 010950) +8.86% (intraday) | ❌ Not a liquidation tape |
| Defense | Korea Aerospace Industries (KRX: 047810) +5.41%, LIG Defense & Aerospace +5.61% (intraday) | ❌ That’s rotation, not panic |
| Flows (KOSPI) | Retail +₩3.884T (~$2.58B); foreigners -₩1.706T (~$1.14B), institutions -₩2.222T (~$1.48B) | ❌ A handoff, not a stampede for the exit |
| Domestic 2x SK hynix ETFs | -32% (1Q, KIWOOM, ACE, TIGER — Korean single-stock leveraged ETF brands) | ⚠️ The amplifier |
Refiner and defense moves are intraday quotes from Monday’s session; treat closing prints as provisional.
Here’s the turn. If the market as a whole had broken, the currency would have gone first. Currencies are where a country-level scare shows up first, because foreign money has to sell the currency to leave — and Korea imports roughly 70% of its crude from the Middle East (per Korea National Oil Corporation figures), so a genuine Hormuz closure is a Korea macro event, not a chip-sector event. The won should have moved 20 won, not 2. A 0.13% move means the FX market priced Monday as an energy shock, not a Korea systemic shock. And that structural concentration — two names carrying an entire index — is exactly the fault line we mapped in our breakdown of what you actually own when you buy EWY, the main US-listed Korea ETF. (Short version: buy EWY and you are, to a surprising degree, buying Samsung and SK hynix.)
Where the conventional wisdom is right
The geopolitics is real. Hormuz traffic has collapsed and crude is up over 4%. For an economy that imports its energy and exports its manufactured goods, that lands on margins — airlines, chemicals, power costs — with a lag.
The valuation retrace is real, too. SK hynix was one session past its Nasdaq debut as an ADR — an American Depositary Receipt, a US-listed certificate representing shares held abroad — which closed at $168.01 on July 10, up about 13% on day one. It traded that week under the temporary ticker SKHYV, with the switch to the permanent SKHY reported for July 14. (Each ADR represents a fraction of a Seoul-listed share, so the dollar and won prices in this piece are not directly comparable.) A rally with that slope invites a pullback whether or not anything fundamental changes.
And the “peak-out” debate — Korea’s term for a cycle top — is legitimate. Memory is a cyclical industry. Nothing in Monday’s tape proves the cycle-top argument wrong.
Where it’s wrong
1. “The market” didn’t break — half the index did. Samsung plus SK hynix were 54.76% of KOSPI market cap in mid-June, and more after July’s run. Mark those two down 10.7% and 15.4% and, on their mid-June weights, the arithmetic alone accounts for roughly seven of the index’s nine points. Everything else is hidden behind them — including the refiners that were up double digits.
That is not the same as saying the rest of Korea was fine. Back out the two chip names and the remaining ~45% of the index still fell on the order of 4% — a bad day, just not a systemic one. The claim here is narrower than “only two stocks fell”: it is that two stocks turned a bad day into a 9% day.
2. Nothing that broke on Monday was demand data. What printed was price and flow. Not shipments, not contract prices, not order books. Samsung’s July 7 preliminary operating profit was a record ₩89.4 trillion (~$60B). No HBM cancellations have been disclosed. You cannot read a demand inflection off a tape that contains no demand data.
3. What broke was positioning. Domestic 2x single-stock SK hynix ETFs — funds engineered to return double the stock’s daily move, which is also why they fall roughly twice as fast — dropped about 32% in a session, and margin calls followed: leveraged holders were forced to sell into a falling tape, which pushed it lower, which forced more selling. Retail absorbed ₩3.884 trillion (~$2.58B) — almost exactly the ₩3.928 trillion (~$2.61B) foreigners and institutions threw overboard. That’s a leverage unwind with a buyer of last resort, not a demand verdict.
4. A new amplifier switched on the same day. On July 13-14, New York began listing leveraged and inverse ETFs on the SK hynix ADR: two from Leverage Shares — SKHX (2x the daily move, long) and SKHZ (-1x, i.e. short) — and one from ProShares, SKHU (2x long). That’s per issuer notices; confirm tickers and listing dates with the issuers before trading. Seoul’s ordinary shares and New York’s ADR are now wired together through two-way leveraged products. Volatility can now travel that wire in both directions — and the wire opened for business on the day the stock fell 15%.
🎩 Under the Gat — The index fell 9% and the won moved two won. Bonds didn’t panic. Refiners didn’t panic. What broke wasn’t the market — it was the two stocks that pretend to be the market. When an American reads “Korea crashed,” the thing that actually happened is closer to “Korea’s two chip stocks got margin-called.”
The verdict
Half true.
True: the geopolitical shock and the valuation retrace both happened.
False: “AI demand is cracking” is a conclusion Monday’s data cannot carry. What broke was leveraged positioning, and what amplified it was index concentration.
Not yet known: the data that settles the argument arrives on or around July 22 — SK hynix’s Q2 results. If demand really is rolling over, it shows up there in shipments and pricing. If this was positioning, the earnings should hold up and only the price was hurt. — a view, not advice.
So what do you actually watch?
Margin balances and leveraged-ETF assets. Whether the unwind is finished is a question the balances answer, not the price. If money flows straight back into a 2x product that just fell 32%, the amplifier has been reloaded.
The 1,500 line on USD/KRW. The currency stayed calm on Monday. If that line goes, the story changes character — from a stock problem to a country problem.
July 22, SK hynix Q2. The mechanics and the traps of that print — starting with the fact that this company publishes no preliminary number — are laid out in our guide to SKHY’s first earnings.
Whether the energy rotation sticks. A one-day headline trade or a new oil regime? Brent is at $78.82 — watch whether it takes out $80 and stays there.
The pattern, not the day. Last week already gave Seoul two breaks — a circuit breaker on July 7 and, on July 10, a buy-side sidecar (a five-minute freeze on program trading, the milder cousin of a circuit breaker) — which we logged in the weekly scoreboard. Monday makes seven circuit breakers this year; sidecars are counted separately. In a market this concentrated, the halts are the structure.
🎩 Under the Gat — An American investor’s reflex is “index down 9% = systemic risk.” In Korea that reflex misfires constantly. Here you have to read the composition of the index before you read the index. That’s the price of admission to this market — and the edge you get for understanding it. — a view, not advice.
FAQ
Q. Is the AI trade collapsing?
Monday’s Korean data can’t carry that conclusion. Prices and leveraged positions broke; demand data — shipments, contracts, capex — did not. Samsung’s July 7 preliminary operating profit was a record ₩89.4T (~$60B), and no HBM cancellations have been disclosed. The next data that can settle it is SK hynix’s Q2 print, expected on or around July 22.
Q. The KOSPI fell 8.95% — why didn’t the won move?
Because the FX market read the day as an energy and single-stock shock, not a Korea systemic event. USD/KRW closed the onshore session at 1,503.4, up 2.0 won (+0.13%). In a true broad risk-off, the currency usually goes first — foreign money has to sell the currency to leave.
Q. Seven circuit breakers this year? What about the US?
Monday’s Level 1 halt (13:28 Seoul time) was 2026’s seventh, following a sell-side sidecar at 10:34. US market-wide circuit breakers (S&P 500 at -7%, -13%, -20%) haven’t triggered since March 2020. That gap is as much about index structure as about panic — Korea’s index is far more concentrated in a few mega-caps.
Q. Did anything actually go up?
Yes. On the Hormuz escalation, refiners (SK Innovation, S-Oil) and defense names (Korea Aerospace Industries, LIG Defense & Aerospace) rose — some by double digits intraday — while the index fell nearly 9%. Shipping names rallied too, per Korean market reports. The index line alone hides that.
Q. Can I use leverage on SK hynix from a US account now?
From July 13-14, leveraged and inverse ETFs on the SK hynix ADR are reported to be listing in New York: Leverage Shares SKHX (2x long) and SKHZ (-1x, short), plus ProShares SKHU (2x long) — confirm tickers and listing dates with the issuers before trading. They track daily returns, so in volatile stretches they diverge hard from the underlying — Korea’s own 2x SK hynix ETFs fell about 32% on Monday. Not financial advice.
Sources
- KED Global — Korean stocks suffer another Black Monday as Kospi slips below 7,000 (July 13, 2026)
- Newspim — KOSPI closes at 6,806.93, -8.95% (July 13, 2026) · Edaily — seventh circuit breaker of 2026
- FN News — 2x SK hynix leveraged ETFs fall ~32%; flows and margin calls (July 13, 2026)
- Etoday — refiners surge on renewed US-Iran strikes: SK Innovation, S-Oil (July 13, 2026, intraday)
- Al Jazeera — Brent $78.82 as US and Iran trade attacks over the Strait of Hormuz (July 13, 2026)
- KRX — index and market-cap data; Seoul FX market — USD/KRW onshore close 1,503.4 (July 13, 2026)
- Samsung Electronics — Q2 2026 preliminary results (July 7, 2026): operating profit ₩89.4T
— Mr. Gat 🐂
This is not financial advice. Figures are as of the July 13, 2026 Seoul session; refiner and defense moves are intraday quotes and should be rechecked against closing data. Tickers, listing dates and index weights should be verified against KRX and issuer notices as of your reading date. FX conversions use USD/KRW 1,503.4 (July 13, 2026 onshore close). The author does not recommend buying or selling any security.
Frequently Asked Questions
Is the AI trade collapsing?
Monday’s Korean data does not support that conclusion on its own. Prices and leveraged positions broke; demand data — shipments, contracts, capex — has not. Samsung’s July 7 preliminary operating profit was a record ₩89.4 trillion (~$60B), and no HBM order cancellations have been disclosed. The next data point that can settle the argument is SK hynix’s Q2 result, expected on or around July 22. This is not financial advice.
Why did the KOSPI fall 8.95% while the won barely moved?
Because the currency market read the day as an energy and single-stock shock rather than a Korea systemic event. USD/KRW closed the onshore session at 1,503.4, up just 2.0 won (+0.13%). In a genuine broad risk-off, the won typically moves first and hardest.
How many circuit breakers has Korea had this year?
Monday’s Level 1 halt (13:28 KST) was the seventh of 2026, after a sell-side sidecar at 10:34. For comparison, US market-wide circuit breakers (S&P 500 at -7%, -13%, -20%) have not been triggered since March 2020. The gap says as much about index structure as about panic.
Did anything go up on Black Monday?
Yes. With the Strait of Hormuz escalating, refiners (SK Innovation, S-Oil) and defense names (Korea Aerospace Industries, LIG Defense & Aerospace) rose — some by double digits intraday — even as the index fell nearly 9%. Shipping names also rallied, per Korean market reports. Reading the index line alone hides that rotation.
Can US investors use leverage on SK hynix now?
From July 13-14, leveraged and inverse ETFs tied to the SK hynix ADR are reported to be listing in New York — Leverage Shares 2x Long (SKHX) and 1x Short (SKHZ), and a ProShares 2x product (SKHU), per issuer notices; confirm tickers and listing dates with the issuers before trading. These track daily returns, so in volatile stretches they diverge sharply from the underlying: Korea’s own 2x SK hynix ETFs fell about 32% on Monday. This is not financial advice.
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.