SK Hynix’s Limit-Up Rebound: Vincent’s Take on Sampro TV

The Sampro TV (삼프로TV) weekend interview features Vincent from Hana Securities (하나증권), reflecting on an “unprecedented” July for Korean equities and what SK Hynix’s (SK 하이닉스) violent moves signal. The conversation was recorded on July 31, 2026, just before the market opened, after SK Hynix had effectively halved from roughly 2.64 million won at end-June to about 1.359 million won, before a sharp rebound.

According to Vincent, even with his 35 years of investing experience, the combination of leverage, panic selling, and policy shocks produced one of the worst monthly drawdowns since the IMF crisis. The host underscores that SK Hynix had briefly been Korea’s largest stock by market cap, making its one‑month collapse particularly shocking.

Vincent openly admits he “got it wrong,” having stressed that pullbacks should be bought and under‑emphasized macro risks. He sees SK Hynix’s late‑July limit‑up move and Chairman Chey Tae‑won’s personal share purchase (around 4.8 billion won) as symbolic: insiders are stepping in after forced selling by margined retail, institutions, and foreign investors. The biggest argument Vincent makes is that, despite the trauma and overhang, the underlying earnings power and valuation of Korea’s leading memory makers remain compelling.

Vincent’s central claim is that Korean equities in general, and the large memory makers in particular, look fundamentally undervalued relative to their earnings outlook as of late July 2026. He frames his view explicitly as medium‑ to long‑term, even as sentiment and positioning have flipped from euphoria to despair.

On the market overall, Vincent notes that the KOSPI is trading at about 4x forward price‑to‑earnings, implying that “in roughly four years” of projected aggregate earnings you could theoretically buy the entire market cap. He highlights that Samsung Electronics (삼성전자) trades at roughly 3.7x forward earnings and SK Hynix at about 3.3x.

At the same time, Vincent argues that the memory industry is trying to escape its historical boom‑bust “cyclical” character. He points to long‑term agreements (LTA) with hyperscale customers, as described on recent Q2 conference calls by Samsung Electronics and SK Hynix, as evidence of a shift toward more stable, contracted earnings. In his view, these structures, together with AI‑driven demand for high‑bandwidth memory (HBM), support a thesis that the sector may be moving into a “new era” rather than just another classic peak in a cycle.

To back his constructive stance, Vincent walks through several datapoints he views as supportive of Korea’s equity and memory outlook, even after July’s crash.

On the macro side, he notes that Korean exports surpassed 100 billion dollars in June on a monthly basis, and that preliminary 10‑day and 20‑day export figures for July are also strong. He references the OECD’s leading indicator for Korea, which he says remains aligned with robust export trends, and the U.S. ISM new orders data, which he describes as a statistically leading series for Korean exports with “no clear deterioration” yet.

On micro fundamentals, Vincent stresses that Korean corporates’ earnings per share are on a steep upward trajectory, and he reiterates his house forecast that SK Hynix’s operating profit could reach around 82 trillion won in the second half (he clarifies this is a forecast and much higher than the first half). He emphasizes that prior cyclical peaks were accompanied by ballooning inventories, whereas SK Hynix’s Q2 inventory‑to‑sales ratio is falling; his chart inverts the axis to show inventory drawdown even as share prices collapse.

Most importantly for his structural thesis, Vincent cites Q2 conference call commentary: Samsung Electronics claims to have already signed LTAs with five global customers and to be nearing completion with another five, with typical contract durations of five years and annual price resets. SK Hynix, he says, has LTAs with about 10 customers. Both companies aim to push LTA coverage to 60–70% of volume. To him, this signals an attempt to secure multi‑year earnings visibility tied to AI‑related products like HBM and high‑value DRAM.

Despite his bullish bias, Vincent spends considerable time on the risks that have just played out—and that could still cap the recovery. He characterizes July as a period where “fear overwhelmed logic,” with retail investors heavily exposed to margin credit and leverage products. Client deposits, he notes, fell from about 140 trillion won to the low‑100‑trillion range in roughly a month, and credit balances were hit by forced liquidations as prices slid.

He also acknowledges the huge technical overhang on the KOSPI. Above roughly the 7,000‑point area, he shows a dense band of trading volume, meaning many investors are now trapped at higher levels and may sell on any rebound. He calls this a “massive supply zone” that will require time and patience to clear, arguing that only time and repeated tests can rebuild confidence and drain this overhead supply.

Vincent further concedes that analysts, including himself, may be stuck in what he calls a “value trap”: trained to call stocks cheap at low multiples, they might underestimate the power of liquidity, policy, and positioning to keep valuations depressed. Some brokerages have already cut target prices on Samsung Electronics and SK Hynix, while others—including his own house—maintain or even raise them; he cites his firm’s 3.6 million‑won target for SK Hynix as unchanged. He warns that if investors cling only to screens of low P/E without respecting supply, rates, and foreign flows, they risk repeating July’s pain.

Looking forward from the July 31, 2026 vantage point, Vincent sketches out several timelines and catalysts he believes investors should monitor rather than the day‑to‑day price swings.

First, he analyzes past KOSPI drawdowns of more than 30% from peak, which he says have occurred only about six times in 46 years. On average, it took roughly 1,148 trading days (about three years) to reclaim previous highs. However, because the latest run‑up was unusually steep, he recalculates using the most recent ascent speed and gets a faster potential recovery of about 150 trading days (around five to six months). In an extremely optimistic V‑shaped rebound scenario, his math suggests about 45 trading days.

Based on this, Vincent proposes a “square‑root‑shaped” path: a sharp rebound in August that retraces perhaps around half of July’s decline, followed by a period of range‑bound consolidation into late August through November. He ties that consolidation to heavy overhead supply and to macro events: the U.S. midterm elections and a planned U.S.–China meeting in Washington in September, which he says could influence tech and semiconductor geopolitics.

On China’s CXMT (ChangXin Memory Technologies), Vincent notes that its DRAM market share is around 8%, far below the combined 67% of Samsung Electronics and SK Hynix, and that its yield and product quality are lower. Yet CXMT’s market capitalization, he states, is about 712 trillion won, compared with approximately 2,100 trillion for TSMC, 1,400 trillion for Samsung Electronics, and around 100 trillion for SK Hynix. He views CXMT as richly valued and argues that the recent rush of foreign and hedge‑fund capital into China—especially around CXMT’s IPO—hit Korean semiconductors through portfolio reallocation rather than fundamentals. Vincent expects the valuation and flow shock from CXMT’s listing to fade over time.

Finally, he advises watching big‑tech capex and financing conditions. While U.S. hyperscalers have pledged to “borrow to invest” in AI infrastructure—initially seen as a positive for memory demand—he notes that their free cash flow has turned negative, oil and inflation expectations have ticked higher, and long yields have risen, raising questions about how sustainable that borrowing will be. In his view, if financing markets continue to fund AI capex and LTA volumes hold, it would validate his non‑cyclical thesis; if not, bears arguing that “it’s still just a cycle” may win the debate.

In his closing remarks, Vincent addresses retail investors directly, stressing that the psychological damage from July’s crash is real and that volatility in August could remain high. He likens the environment to a typhoon: investors should “fasten their seatbelts,” reduce over‑leverage, and anchor themselves to clear personal criteria and fundamentals rather than pure sentiment.

He reiterates that companies that truly earn money will eventually be re‑rated, even if the journey is slow and uneven. As an example of long‑term compounding, the host recalls that Samsung Electronics traded near 30,000 won (old 5,000‑won par value, pre‑splits) in early 1998 at the depth of the IMF crisis, while its effective price now—after splits and compounding—is equivalent to roughly 1 million won or more, implying over 300x price appreciation before dividends. Yet he notes that very few individual investors actually captured that full move, underscoring how difficult long‑term holding is.

The host also highlights Chey Tae‑won’s historic and recent SK Hynix purchases as a kind of multi‑decade “bet” whose outcome may only be clear ten years from now. Vincent closes by saying he still looks “upward when everyone is looking down,” but accepts that only time and discipline—not bold forecasts—will determine whether this latest SK Hynix surge marks the start of a durable new leg in Korea’s semiconductor story.