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.
In June 2026, Korea’s market served up two headlines that look like they can’t both be true. On June 18, the KOSPI closed above 9,000 for the first time ever (around 9,064), after SK Hynix shipped next-generation HBM4E 12-layer memory samples to key AI customers. (Verify the exact close.) Days later, another headline read: foreign ownership of Samsung Electronics and SK Hynix just hit a one-year low, with foreigners net sellers of Korean stocks by the tens of billions of dollars this year. Record high — and foreigners heading for the door. Both are real.
The short answer: it’s one machine, not two stories
Foreigners selling “Korea” and foreigners buying “Korean chips” is not a contradiction — it’s two outputs of the same machine. As Korea’s weight in global and emerging-market benchmarks ballooned, many foreign funds were pushed past their risk limits and had to trim Korea as a whole — not because they dislike it, but because the rules say so. That’s forced selling. At the same time, the conviction that remains is pouring more deeply into a narrow slice: AI memory. So the index hits 9,000 and foreign ownership of the chip giants hits a low — at the same time. What matters isn’t “are foreigners buying or selling?” It’s “where are they concentrating, and what are they emptying out?” This is not financial advice.
🎩 Under the Gat — When you see “foreigners dump Korea” next to “KOSPI hits 9,000,” it feels like one must be a lie. Neither is. Fund managers aren’t selling because they hate Korea — they’re selling because Korea got too big for their rulebook. And the real conviction? It’s digging deeper into just two chip names. Watch the concentration, not the index.
Lens 1 — the “forced seller”: Korea got too big
The first lens is foreign money that’s tied to a benchmark. As Korean stocks ran hard over a year, Korea’s weight — especially Samsung and SK Hynix — surged inside MSCI Emerging Markets and global indices. Many active managers cap how much any single country or stock can be in a portfolio (a risk limit). When the weight blows past the cap, they have to sell some — not out of dislike, but out of policy. One market account called it “essentially forced selling.” (Verify; Seoul Economic Daily.)
The numbers: by Goldman’s tally, foreigners net sold roughly $62B of Korean stocks as of late May, and trimmed about $22B more since. The broad Korea ETF, EWY (iShares MSCI South Korea), bled around $1.01B in the first week of May alone. (Verify.) Read this way, the “selling” is portfolio management, not a bearish call — closer to “Korea exceeded my limit” than “I’m out on Korea.”
🎩 Under the Gat — Here’s what outsiders get wrong most often: net selling does not mean the bull market is over. When an index weight explodes, funds trim even their winning positions to stay inside the rules. The headline says “foreign sell-off.” What’s actually happening in the engine room is rebalancing.
Lens 2 — the “conviction buyer”: chips get bought harder
The second lens is the opposite flow hiding inside that same word, “foreigners.” Money leaving the broad KOSPI is crowding into a few AI-memory names. Over four sessions from June 11, foreigners net-bought roughly 3.1 trillion won of SK Hynix, and of the net buying that pushed the KOSPI above 9,000 on June 18, Samsung Electronics was the single largest chunk. (Verify; ConnectMoney/Bloomingbit.)
The ETF flows tell the same story: outflows from the broad EWY, but roughly $1.95B of inflows into a DRAM-focused ETF (e.g., Roundhill Active DRAM) in that first week of May. (Verify.) Trim the “surface” of Korea; bet harder on the single “point” of AI memory. In this lens, foreigners aren’t leaving Korea — they’re narrowing. Their conviction simply shrank from “Korea” to “Korea’s AI memory.”
The anchor for US readers: Korea’s “Magnificent 2” vs. America’s “Magnificent 7”
This structure isn’t strange to a US investor. Since 2023, a handful of AI leaders — the Magnificent 7 — have carried the US market while breadth (how many stocks actually participate) stayed weak. Korea is the more extreme version: not seven stocks, but effectively two.
| Korea (KOSPI) | US (S&P 500) | |
|---|---|---|
| Concentration core | Samsung + SK Hynix (“Magnificent 2”) | Magnificent 7 (Apple, Nvidia, etc.) |
| Weight in the index | Two stocks ≈ ~48% (verify) | Seven stocks ≈ 30%+ (varies) (verify) |
| Investor crowding | Avoid broad market + pile into AI memory | Weak broad market + pile into a few AI names |
| Core risk | Two stocks hold the index’s fate | Spread across seven (less extreme) |
| Decisive difference | One notch more extreme — tied to a single memory cycle | Big Tech is diversified (search, ads, cloud, AI) |
Not a perfect parallel — the US Magnificent 7 run diversified businesses, while Korea’s two giants are both exposed to the same memory cycle. The quality of the concentration is different.
🎩 Under the Gat — Wall Street calls this “narrow breadth.” In Korea it’s simpler — two daejang-ju (대장주), the “general” stocks that lead their sector, are half the index. You think you’re buying the market; you’re really buying two memory stocks. Brightest when things are good, most painful when they turn.
Both sides, fairly — bull vs. bear
The bull case. The chronic Korea discount — the long undervaluation of Korean stocks blamed on governance and geopolitics — is finally unwinding into a re-rating, helped by the government’s Value-up reform push and real earnings. Bulls argue HBM is structural demand, not a passing cycle, and point to Goldman raising its KOSPI target to 12,000 (from 9,000) on June 3. (Verify.)
The bear case. With the index lashed to two stocks, if those two shake, the whole market shakes. In early June, when Broadcom’s AI-chip guidance came in below expectations, the KOSPI dropped sharply intraday, tripped a circuit breaker (an automatic trading halt), and Samsung and SK Hynix fell together. (Verify; TradingKey.) CNBC’s warning to respect memory’s boom-bust cycle sits in the same camp.
Layer a flows lens on top: the stock foreigners are trimming from the broad market is being absorbed by the donghak gaemi (동학개미) — Korea’s retail army, literally the “righteous ant army” — who treat Samsung and SK Hynix as kookminju (국민주), “national stocks” so widely held that owning them feels close to patriotism. The fundamentals are real (Korea’s ppalli-ppalli, “hurry-hurry,” execution speed shipped HBM4E samples first), but the cast of buyers is changing.
🎩 Under the Gat — Separate the engine from the fuel. The engine — HBM and AI-memory earnings — is real. But the changing of the fuel, retail stepping in where foreigners are forced to step out, is flows, not fundamentals. Confuse the two and you’re not buying the boom — you’re buying the concentration trap. I don’t get to decide who’s right. The memory cycle and margins do.
The Korean texture: who catches what the foreigners drop
Who’s holding up the market that foreigners are trimming for rulebook reasons? The donghak gaemi. To them, Samsung and SK Hynix aren’t just tickers — they’re kookminju, national stocks where tech leadership feels almost like national pride. Retail catching the record-high stock that foreigners let go is exactly the kind of local tension an outsider has to understand to read Korean flows at all.
Risks worth naming (things to watch, not predictions)
- Two-stock dependence. With Samsung and SK Hynix near half the index, the whole market hangs on HBM pricing and the memory rivalry (Micron, Samsung’s catch-up).
- A concentration shock. A single supplier’s weak guidance (see Broadcom) can drag both giants — and the index — down fast. (Verify.)
- Foreign flows and FX. Continued benchmark-driven selling plus a moving won/dollar rate changes returns for foreign-based investors.
- “Index” that isn’t. A broad Korea ETF is now close to a concentrated memory bet — check the holdings before assuming you’re diversified.
The bottom line: foreigners selling Korea and foreigners buying Korean chips are the same machine seen from two sides. Tell the forced selling apart from the conviction buying — and the broad index apart from the two stocks inside it — and the picture stops contradicting itself.
— Mr. Gat 🐂
This is not financial advice. Past performance does not guarantee future results. Figures marked “verify” are drawn from press reports as of June 2026 and should be confirmed against primary sources before you rely on them.
Frequently Asked Questions
If foreigners are net sellers, why is the KOSPI at a record high?
A large share of the foreign selling isn’t a bearish bet — it’s forced selling. As Korea’s weight in global and emerging-market benchmarks surged, many funds had to trim “Korea” to stay within risk limits. Meanwhile domestic retail investors and AI-memory-focused money pushed the index to records. This is not financial advice.
But didn’t foreigners net-buy SK Hynix?
Yes — and that’s the whole point. The same word “foreigners” covers two opposite flows: trimming the broad market while concentrating into AI memory. Foreign investors net-bought roughly 3.1 trillion won of SK Hynix over four sessions in mid-June 2026, even as broad Korea exposure shrank. (Verify exact figures.)
Does buying EWY get me this chip rally?
Partly. EWY (iShares MSCI South Korea) is heavy in Samsung and SK Hynix but is still a broad Korea index. In 2026, money flowed out of broad Korea ETFs while DRAM-focused ETFs drew inflows. If you want pure chip exposure, check the holdings first. This is not financial advice.
Why is it risky that Samsung and SK Hynix are nearly half the index?
Because if those two wobble, the whole index wobbles. In early June 2026 a weak Broadcom AI-chip guidance dragged both stocks down and the KOSPI fell sharply intraday, tripping a circuit breaker. Concentration amplifies both the upside and the downside. (Verify.)
So should I buy Korean chip stocks now?
This is not financial advice. But the axes to judge are clear — the fundamentals (HBM demand and margins), the flows (foreign forced selling vs. retail buying), and the concentration risk (a two-stock index). Looking at them separately is how you tell a boom from a concentration trap.
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.