Invest1 distinct publisher3 min readUpdated
Korean retail investors put about $840 million into SK Hynix's U.S. depositary receipts in July, part of $4.5 billion of net U.S. buying. The receipts have traded around 10% above the Seoul shares.
The Investor · Invest desk

Compiled by The InvestorSomething wrong?How this is made
Korean retail investors net bought roughly $4.5 billion of U.S. stocks in July, according to Korea Securities Depository data, and about $840 million of that went into the U.S.-listed depositary receipts of SK Hynix, a company they can buy directly at home [1][2]. Those receipts have recently traded about 10% above the Korean shares, according to Owen Lamont, senior vice president at Acadian Asset Management, which makes this flow something other than a reallocation decision [3].
The concentration is the first thing worth sitting with. One Korean company's ADRs absorbed close to a fifth of the entire month's net U.S. buying by Korean retail, and ranked as the second most net-purchased U.S. security [2][1]. If the whole tranche had been transacted at a 10% premium, the arithmetic implies about $76 million paid above the home-market value of the same shares, for a security that is also more volatile than the Seoul line [2][4]. Lamont's assessment was blunt: "That's absolutely crazy," he said, adding that "there's no reason for a Korean investor to buy ADRs of Korean stocks in the U.S." [5] He called such price discrepancies unusual and a possible warning sign of speculative excess, "a symptom of the bubble," pointing to similar dislocations involving Taiwanese and Indian companies around the dot-com boom [6].
The surrounding behavior fits. Four of the ten most net-purchased U.S. securities in July were leveraged products, led by the Direxion Daily Semiconductor Bull 3X Shares ETF, with ProShares UltraPro QQQ fourth and ProShares Ultra QQQ sixth [7]. Ultra QQQ ranked seventh this month [8]. Meanwhile Korean retail net sold domestic stocks for most of last week even as the benchmark entered bull market territory, and foreign investors turned net buyers, per Korea Exchange data [9]. Domestic margin loan balances stood near 37 trillion won, about $26 billion, at the end of June, then fell to 27 trillion won earlier this month, the lowest level this year [10] - a roughly 27% unwind [3].
So the leverage came out at home and the same bet went back on abroad. Phillip Wool of Rayliant Global Advisors said the purchases are "largely shares tied up in the same AI hardware theme that's been selling off in the local market" [11]. Jung In Yun of Fibonacci Asset Management said traders bruised by Korean semiconductor and leveraged-ETF losses may be moving to U.S. AI names they see as higher quality or more liquid, and "are not necessarily reducing their exposure to the AI theme" but "changing the geographical vehicle through which they express the same view" [12]. Lamont noted July's buying was strong but not unprecedented, and that the interesting part is that it rose while the Korean market was plunging [13][14].
For anyone carrying Korea exposure, the practical read is that the domestic retail bid is not returning, it is being exported, and it is being exported at a cost. On spillover, the two views diverge: Wool sees little risk to overall U.S. volatility given institutional dominance of turnover, while Lamont expects distortion in individual names and retail-favored corners, citing the late-2024 rush into U.S. quantum stocks and leveraged ETF proliferation across Korea, Hong Kong and the U.S. [15][16].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
The U.S. receipts have traded at a premium to the Korean shares, which Owen Lamont, senior vice president of Acadian Asset Management, said was about 10% recently.
The U.S.-listed receipts are also exhibiting greater volatility than the Korean shares.
Lamont said such price discrepancies are unusual and can be a warning sign of speculative excess: "They're a symptom of the bubble," pointing to similar dislocations involving Taiwanese and Indian companies around the dot-com boom.
Phillip Wool, head of research at Rayliant Global Advisors, said that if you parse the data on what Korean investors are buying, "it's largely shares tied up in the same AI hardware theme that's been selling off in the local market."
Jung In Yun, founder of Fibonacci Asset Management, said some traders hurt by losses in Korean semiconductor shares or leveraged ETFs may be shifting to U.S. AI stocks they perceive as higher-quality or more liquid: "They are not necessarily reducing their exposure to the AI theme... They may simply be changing the geographical vehicle through which they express the same view."
Korean retail investors net bought around $4.5 billion in U.S. stocks in July, according to Korea Securities Depository data, a sharp pickup from June and near January's net purchases of $5 billion.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Hard flow data, single outlet, unverified premium
The quantitative spine — monthly net purchases, the SK Hynix ADR figure, leveraged ETF rankings, margin balances, retail versus foreign flows — is attributed to named institutional data providers (Korea Securities Depository, Korea Exchange, Korea Financial Investment Association) and is internally consistent. Weakening it: one publisher carries the cluster, the headline ~10% ADR premium is a single analyst's recollection rather than a published price series, and the market-impact conclusions are unquantified opinion.
Behavior is large and documented, concentration verified
The behavior at issue is measurably widespread: about $4.5 billion of net U.S. buying in a single month with roughly 19% concentrated in one company's ADRs, and leveraged products occupying four of the ten top slots. Adoption is scored on the documented scale of the flows, not on any product uptake; it is held below the top band because composition beyond a few line items is undisclosed and the data is one month of a single agency's tally.
Data solid, bubble framing runs ahead of it
The flows and rankings are as reported, but the interpretive frame — 'absolutely crazy,' 'a symptom of the bubble,' dot-com analogies — is stronger than the supplied evidence. The premium is one recent estimate, no benign explanations (trading hours, FX, brokerage access) are tested, no arbitrage or realized-loss data is offered, and Lamont himself concedes July buying was strong but not unprecedented while Wool doubts any aggregate U.S. impact. Modestly overstated rather than unfounded.
Commentary sourced entirely to asset managers
Every interpretive claim comes from professionals at active investment firms — Acadian Asset Management, Rayliant Global Advisors and Fibonacci Asset Management — whose businesses involve positioning on Asian and global equities and on market mispricing; the strongest 'bubble' language comes from the manager most identified with mispricing research. The publisher itself has audience incentives around vivid market framing. No corporate, brokerage or regulatory party with a stake in the ADR flow is quoted, and the underlying data agencies are neutral, which caps the score mid-range.
Facts dependable, conclusions soft
Confidence is moderate: the numeric core is agency-sourced and arithmetically checkable, so the descriptive claims are dependable, but the cluster has one publisher, the premium is unverified, and the consequential claims about bubble dynamics and volatility transmission are hedged single-analyst forecasts marked insufficient here.
invest
Korean retail paid a 10% premium to own SK hynix in New York instead of Seoul1 distinct publisher
invest
Korean retail money quit the 3x chip trade: $4.42B in July, $106M so far in August1 distinct publisher
invest
SK's chairman warns of worst-ever memory imbalance while a U.S. fab stays under review1 distinct publisher
invest
Five brokers cut JYP targets because two groups' availability is the entire thesis1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 17, 2026