SK Hynix's Rebound Test: Can Record Profits, a 100 Trillion Won Payout, and a $38 Billion Bet Calm Investors?

SK Hynix posts record profits but faces investor doubt over $72B shareholder return, $38B capex, and HBM4 delays. Stock remains 33% below 50-day MA.

The whiplash hitting SK Hynix shareholders this week is a masterclass in market psychology. After a brutal month that erased more than a third of the stock's value, the Seoul-listed memory chip giant is trying to thread a needle: convincing investors it can simultaneously return record cash, absorb a massive capital expenditure program, and fend off a resurgent rival—all while a labor dispute simmers in the background.

The shares have stabilized, trading around 1,430,000 won in recent sessions, but the recovery is fragile. The stock remains roughly 33% below its 50-day moving average, and the Relative Strength Index sits at 38.8, flirting with oversold territory. For a company that just posted its best quarter ever, the market's skepticism is striking.

The Numbers That Should Have Been Enough

The financial results SK Hynix unveiled in late July were, on their face, spectacular. Operating profit surged 557% to 60.54 trillion won, while revenue climbed 257% to 79.32 trillion won. Net income reached 93.92 trillion won—more than thirteen times the prior-year figure—though that headline number was flattered by 63.3 trillion won in gains from the sale of the Kioxia stake, a deal completed in June after a seven-year holding period.

Yet the market punished the stock by 9.6% on the day. The problem wasn't the absolute numbers but the miss against expectations. Analysts had penciled in operating profit of 64 trillion won and revenue of 84 trillion won, according to Reuters. The shortfall was attributed to weaker-than-expected HBM4 shipments, with revenue recognition for those advanced memory chips deferred to later quarters. The company guided to DRAM shipments roughly 10% above the prior quarter in Q3, with a full HBM4 production ramp slated for the second half of the year.

A Carrot, Then a Stick

Management's response to the investor backlash came in two parts. On Friday, the board approved a quarterly dividend of 375 won per share—about $0.27—totaling roughly $198 million, with a record date of August 31 and payment within a month. More significantly, media reports indicate the company is preparing a shareholder return program worth around 100 trillion won (approximately $72.4 billion), combining buybacks and cash dividends, with the buyback component potentially reaching 40 trillion won. Management said it would finalize details in the third quarter.

The market's reaction was telling: the stock fell another 4.88% on the announcement. Investors who had watched SK Hynix generate record profits while delivering lackluster shareholder returns wanted more than promises.

Then came the capex bombshell. The same Friday, the board approved investments totaling 54.3 trillion won (roughly $38.3 billion) in two new fabrication plants. The larger tranche—35.2 trillion won—goes to the "Y2" DRAM fab in the Yongin semiconductor cluster, with construction starting in July 2027 and the first clean room slated for June 2029. The remaining 19.1 trillion won funds the "M17" NAND fab in Cheongju, occupying about 680,000 square meters, with groundbreaking in February next year and the first clean room module opening in December 2028. Industry analyst firm Omdia projects annual DRAM and NAND growth of around 19% through 2030 as a result.

The stock dropped another 4.9% on that news. Investors seemed to be asking a simple question: if you're spending $38 billion, how much will actually come back to us?

The Analyst Split

Wall Street is divided on what happens next. Morgan Stanley's Sean Kim—the analyst who famously warned "Memory, Winter Is Coming" in 2021—now sees the correction as over. In a report dated August 6, he reaffirmed a price target of 2.6 million won for SK Hynix and raised 2026 earnings estimates by 13%. His evidence: foreign selling is decelerating, from outflows of roughly $30 billion in June and $6.2 billion in July to $4.3 billion so far in August. Notably, he cut Samsung Electronics' earnings estimate by 10% while holding that target at 375,000 won.

Cantor Fitzgerald took a different tack, initiating coverage on August 4 with an "Overweight" rating and a $300 price target on the US-listed ADRs—the highest on Wall Street, versus a range of $200 to $250 from other firms. The ADRs had closed at $142.72 before that call.

Labor and the Suwon Threat

Two overhangs complicate the bull case. On the labor front, SK Hynix and its union have gone through five negotiation rounds without resolving a plan to pay part of bonuses in shares. More than 3,500 employees—about 10% of the workforce—want to form a separate union. The dispute carries a real risk: delaying the HBM4 ramp-up, whose mass production began in the second quarter and whose expansion is planned for the second half.

Meanwhile, Samsung Electronics is closing the gap. The Suwon-based rival claims HBM4 yields of around 80%, four months ahead of schedule, with HBM4E reliability-test yields above 70%. UBS projects Samsung will capture 41% of the HBM market next year, versus 39% for SK Hynix.

The China Question

Adding another layer of complexity, Bloomberg reports that SK Hynix is exploring options for its Chongqing facility, including bringing in an investor. The plant, built in 2013 and handling exclusively back-end packaging and testing, could be valued at around $3 billion in a partial sale. The company is in preliminary talks with Chinese funds and industry players but might retain a minority stake. SK Hynix has not officially commented.

A Curious Technical Footnote

Thursday's trading produced an oddity worth noting: at the alternative Nextrade exchange, the stock hit its 30% daily limit down in thin pre-market trading before recovering to a loss of roughly 2% during the 50-minute session. The move appeared to be a technical artifact rather than a fundamental signal.

For now, the stock's trajectory hinges on whether management can convert its promises into tangible returns before the next earnings cycle. The company has the profits, the payout plan, and the capacity investments. What it lacks is the market's full confidence—and that, unlike fabs, cannot be built on a fixed timeline.