On the morning of July 2, 2026, Samsung Electronics shares fell 9.06 percent to close at 286,000 won, dragged down alongside a 14.57 percent plunge in SK Hynix to 2,187,000 won, as a tech-stock selloff that started on the Nasdaq rolled through Seoul. Underlying chip export data held up fine that day, analysts told CNBC — the move looked more like foreign investors repositioning than any change in Samsung's fundamentals. But the reason a single bad session in two stocks could shake Korea's entire market is itself the story: Samsung and SK Hynix's combined weight in the KOSPI index has surged from roughly a quarter at the end of 2025 to more than half by June, hitting close to 60 percent including preferred shares, according to BusinessKorea and Bloomberg. Goldman Sachs has warned that one more percentage point of concentration could trigger forced selling of roughly $2 billion by foreign funds bound by US diversification rules. Korea's stock market, in other words, has become a leveraged bet on whether two companies win the AI memory-chip race.
Samsung earned that gravitational pull the hard way, over four decades of betting on memory chips before anyone called it an AI story. The company posted record quarterly revenue of 133.9 trillion won in the first quarter of 2026 — roughly $100 billion — with operating profit of 57.2 trillion won, an eightfold jump from a year earlier, driven almost entirely by its chip division. It employs more than 259,000 people across 76 countries, with about 125,000 in South Korea, and it remains the world's largest maker of DRAM (38 percent global share) and NAND flash memory (29 percent), the chips that sit inside nearly every phone, server and laptop sold anywhere. Samsung Group's revenue was equal to 22.4 percent of South Korea's entire GDP as of the most recent widely cited comparison, from 2022; affiliates across the group account for roughly a fifth of the country's exports. There is no straightforward American or European analogy for what Samsung is to Korea — it is closer to being the country's chip industry, its consumer-electronics industry and a meaningful share of its stock market, all housed under one roof.
A Conglomerate Run on Borrowed Ownership
That roof rests on a structure few outside Korea would recognize as normal corporate governance. The Lee family controls Samsung Electronics not through a majority stake — they own barely more than one and a half percent of the company directly — but through a chain of cross-shareholdings: the family controls Samsung C&T, which owns a large stake in Samsung Life Insurance, which in turn holds a significant stake in Samsung Electronics. It is a structure common to Korea's chaebol conglomerates and long criticized by foreign investors, who hold roughly a third of KOSPI's market cap, for opaque related-party dealings and low dividend payouts relative to profits.
The structure was tested directly this year. In May 2026, the Lee family completed payment of a record 12 trillion won — about $8 billion — in inheritance tax stemming from the 2020 death of chairman Lee Kun-hee, taxed at a 60 percent rate, among the highest inheritance levies in the world. The family financed it through share sales, dividends and securities-backed loans rather than a fire sale of control, and Lee Jae-yong's direct stakes actually rose as a result — in Samsung Electronics, from 0.70 percent to 1.67 percent. Lee himself has spent much of the past decade in and out of courtrooms: convicted in 2017 of bribery and embezzlement tied to a political scandal that helped topple a Korean president, pardoned in 2022, then acquitted on appeal of separate charges of improper merger and accounting fraud, a verdict upheld by South Korea's Supreme Court in July 2025.
Catching Up in the Only Race That Matters
For most of the past two years, Samsung has been chasing rather than leading in the chip category investors care about most: high-bandwidth memory, the stacked chips that feed Nvidia's AI accelerators. SK Hynix reached volume production of HBM3E in March 2024; Samsung's version failed Nvidia's qualification tests that April, failed again in mid-2025, and did not pass until September 2025. In June 2026, SK Hynix overtook Samsung as Korea's most valuable company for the first time in 26 years, built substantially on its HBM lead — some reporting puts SK Hynix's share of Nvidia's next-generation HBM4 allocation at roughly 70 percent. Samsung's foundry business, the unit that manufactures chips designed by other companies, has fared worse still: its global market share slipped to 6.5 percent in the first quarter of 2026, against TSMC's 72.3 percent, roughly eleven times Samsung's size in that segment.
"Samsung is back." — Jun Young-hyun, Samsung Electronics co-CEO, New Year address, January 1, 2026, relaying customer feedback on the company's HBM4 progress
Jun's line, delivered alongside an acknowledgment that Samsung "still had work to do," reflects a genuine turn rather than pure bravado. Samsung began first commercial shipments of HBM4 memory in the first quarter of 2026 for Nvidia's Vera Rubin GPU platform, and the company has said it will expand HBM production capacity by roughly 50 percent this year. Its foundry unit is reportedly in talks with Google, Nvidia, Tesla, AMD and BYD over advanced-node orders, and some analysts expect that business to turn profitable as early as the third quarter of 2026 — a recovery, if it holds, rather than a resolution.
The optimism sits alongside real friction on the ground. In May 2026, more than 47,000 Samsung Electronics workers prepared to strike over stalled wage negotiations — a walkout that would have disrupted global memory supply given Samsung's centrality to the market — before the company and union reached an eleventh-hour deal days before the planned start date. It is the kind of labor dispute a company this large, this fast-growing and this politically entangled with the state can no longer treat as a private matter: a strike at Samsung is a supply-chain event for half the world's electronics makers. Layer on top of that the index-concentration risk that Goldman Sachs has flagged, and a fair reading of Samsung's position in mid-2026 is neither triumphant nor troubled but precarious in a specific, structural way — a company whose fortunes are increasingly inseparable from Korea's own, at a moment when both are betting heavily on a technology cycle that has not finished playing out.
What happens next depends on decisions largely outside Samsung's control: whether Nvidia's roadmap keeps rewarding whichever memory maker ships fastest, whether foreign funds tolerate a Korean index that trades like a two-stock basket, and whether a foundry business built to serve other people's chip designs can convince customers to trust it at the leading edge. Samsung's own history suggests patience — this is a company that spent a decade losing money in memory before it dominated the category. Whether Korea's stock market, and the pension funds and retail investors riding along with it, can afford to wait that long again is a separate question, and one Seoul is now watching as closely as Samsung's balance sheet.