Korea’s trading halts pause the selling, but leave steep losses intact

For investors watching their South Korean shares plunge, the selloff brought two kinds of loss: sharply lower quoted values, and stretches when they could not trade at all. Twenty-minute halts interrupted both major stock markets on October 15, 2026. They did not erase the damage. At 1:17 p.m. Korea Standard Time, the benchmark KOSPI was still down 8.13%. ^{1 36}

The Korea Exchange’s interventions escalated with the selling. Brief morning restrictions on program sell orders gave way to first-level circuit breakers across the KOSDAQ and then the KOSPI around midday, stopping trading in each affected market and associated derivatives. The distinction matters: the exchange first restricted a category of orders, then stopped the markets themselves. ^{9 12}

The fall was severe from the opening bell, not something that appeared suddenly at the midday thresholds. Nor was it a straight descent. Yonhap reported the KOSPI at 5,604.6, down 6.73%, at 11:20 a.m., a partial recovery from an earlier reported reading before selling deepened again. That morning observation could not be independently rechecked. ^{1 3}

Another morning snapshot needs particular caution. Korea JoongAng Daily reported 5,539.51 at 10:30 a.m. alongside a 7.77% decline. Against the established previous close of 6,008.81, that level implies approximately 7.81%. The quoted level and percentage do not reconcile, and neither should be mistaken for the day’s final loss. ^{4 15}

The KOSPI’s morning restraint, known as a sell-side sidecar, watches futures rather than the cash index. Its trigger is a decline of at least 5% for one minute in the KOSPI 200 futures contract with the preceding day’s highest trading volume. The relevant futures price was down 5.27% when the five-minute suspension activated at 9:07:42 a.m. ¹²

The later KOSPI circuit breaker instead requires the cash index to remain at least 8% below its previous close for one minute. That intervention reaches all shares in the market. The complete official numerical trigger provisions for the KOSDAQ mechanisms were not established in the available material; the KOSPI’s detailed rules should not simply be assigned to its counterpart. ^{6 11}

Reopening was also a process, not an instantaneous switch. The KOSDAQ release notice provided for ten minutes of order collection, followed by a single-price auction and then continuous trading. Block trades could execute during order collection. Yonhap’s 12:44 p.m. bulletin confirmed the halt had been lifted. No exact actual KOSPI reopening time was established, although its later afternoon reading was available. ^{10 14 36}

The historical comparison is narrower than “first dual market halts since July.” MoneyToday’s October 15 report established that the simultaneous morning sell-side sidecars were the first since July 29, 51 trading days earlier. That does not establish the same record for the broader midday circuit breakers. ³⁷

Major chipmakers bore heavy losses. Korea JoongAng Daily’s 10:30 a.m. snapshot put Samsung Electronics at 211,000 won, down 9.05%, and SK hynix at 1,267,000 won, down 10.71%. Morning reporting also put foreign investors’ net selling at 1.92 trillion won, approximately $1.4 billion, extending net sales into an eighth consecutive session. Those figures were not independently rechecked and remain intraday reports, not final returns or completed daily flows. ^{4 2}

The selling preceded Thursday. The Wall Street Journal reported that the KOSPI closed approximately 5.5% lower on October 13. On October 14 it fell another 0.44%, closing at 6,008.81, while the KOSDAQ lost 1.57%. Foreign investors sold a net 1.18 trillion won during that completed session. Comparing that full-day flow with Thursday’s morning figure requires keeping their different cutoffs in view. ^{31 15 16}

The concerns surrounding technology shares were substantial, but their individual contributions to the decline have not been measured. Earlier reporting described rising financing costs, stretched valuations and uncertainty over whether expensive AI infrastructure would deliver lasting profits. A concrete example was the collapse of Nvidia-backed data-centre operator Firmus Grid Ltd.’s initial public offering, reported by Bloomberg through the Financial Post. That failed listing does not quantify its effect on Korean share prices. ^{27 29}

Chinese memory-chip competition added another concern. The Guardian reported CXMT’s claim of progress on a new DRAM architecture intended to increase memory density. That is a technology claim, not confirmation of extra commercial supply. Production readiness, capacity, shipment timing and customer qualification were not established. A future memory glut or loss of incumbent producers’ market share remained a risk—not a verified delivery of new chips to customers. ^{25 31}

Reports also cited U.S. inflation and interest-rate worries. The precise U.S. inflation and stock-index figures accompanying the initial accounts could not be independently rechecked. More importantly, a list of plausible pressures is not a calculation of causation: the reporting does not establish how much of Thursday’s Korean decline came from rates, AI financing or Chinese competition. ^{1 2}

Two dramatic online claims remain unverified: that $203 billion in value was wiped out and that the KOSPI stood 40% below its June peak. The former needs matching market-capitalization observations and a currency-conversion basis; the latter needs an established peak and comparison reading. An October 13 report that Korean market value had fallen below 5,000 trillion won supplies neither calculation for October 15. ^{1 20}

The unfinished session is the final boundary on what can be said. Available reporting establishes no October 15 closing level for either index, and therefore no final daily loss or closing-price losing streak. What it does establish is consequential enough: both markets crossed their first-level interruption thresholds, and the latest verified KOSPI reading still showed a loss exceeding 8%. ^{9 36}

SBS, October 15 market halts ¹; Korea Times/Yonhap, morning trading curbs ²; Yonhap, intraday losses ³; Korea JoongAng Daily, October 15 selloff ⁴; CNBC, CoreWeave financing concerns ⁵.

Korea Exchange, KOSPI circuit-breaker rules ⁶; KOSDAQ sidecar rules ⁷; KIND activation criteria ⁸; Digital Times, dual circuit breakers ⁹; Seoul Economic Daily/KOSCOM, KOSDAQ reopening notice ¹⁰; KRX, KOSDAQ trading provisions ¹¹; SIGNAL, KOSPI sidecar notice ¹².

MoneyToday, English account of morning sidecars ¹³; Yonhap, KOSDAQ halt lifted ¹⁴; Digital Today, October 14 completed session ¹⁵; Asia Business Daily, October 14 close ¹⁶; KRX Data Marketplace ¹⁷.

ChosunBiz, October 13 foreign selling ¹⁸; Businesskorea, opening-market report ¹⁹; Seoul Economic Daily, October 13 market value ²⁰; Herald Business, earlier chip-stock losses ²¹; ChosunBiz, October 12 AI concerns ²²; Sisnoo, October 14 market report ²³; Korea Stock Insight, market overview ²⁴.

The Guardian, CXMT and AI-investment concerns ²⁵; Forbes, chip-stock declines ²⁶; Seoul Economic Daily, AI funding fears ²⁷; Bloomberg via Yahoo Finance, AI-trade concerns ²⁸; Bloomberg via Financial Post, Firmus Grid and financing ²⁹; Baystreet, AI market commentary ³⁰.

The Wall Street Journal, CXMT announcement and Asian stocks ³¹; Associated Press, October 13 semiconductor selloff ³²; Seoul Economic Daily, October 14 capitalization report ³³; The Wall Street Journal, alternate article URL ³⁴.

SIGNAL, KOSPI circuit-breaker notice ³⁵; Korea Herald, 1:17 p.m. KOSPI reading ³⁶; MoneyToday, Korean account of morning sidecars ³⁷; SBS Biz, morning program-selling restrictions ³⁸; Yonhap, intraday volatility report ³⁹.