Why is SK Hynix Stock Falling Despite Record Profit: Better Entry in SKHY or Peak-Cycle Trap?

SK Hynix has delivered one of the strangest earnings reactions of the AI boom. Yet, on July 29, its Seoul-listed shares closed roughly 53% below the all-time high. Meanwhile, company's ADR stock is down 33% from its peak. The second-quarter operating profit surged 557% year-on-year to a record ₩60.5 trillion. Revenue climbed 257% yet the stock tanked. 

The apparent contradiction disappears once we separate record results from record expectations. Revenue and operating profit were both about 5% below forecasts. Headline net profit was inflated by investment gains. Investors were also questioning AI spending, rising manufacturing capacity and the lack of a detailed shareholder-return plan.

Even after the correction, SK Hynix stock remains roughly 115% higher in 2026. Let’s break down what actually changed, what remains intact, and why investors accessing SK Hynix through the Nasdaq-listed SKHY ADR need to examine an additional risk that does not appear in the company’s income statement.

Why Did SK Hynix Stock Fall After Record Earnings?

Six factors came together to push SK Hynix stock lower.

1. SK Hynix's Record profit was still below expectations: SK Hynix reported revenue of ₩79.3 trillion against an approximately ₩84 trillion consensus. Operating profit of ₩60.5 trillion was below the roughly ₩64 trillion LSEG SmartEstimate. The misses were approximately 5.6% and 5.4%, respectively.

A 5% miss would normally be manageable. But SK Hynix had been priced for repeated upgrades after a spectacular rally. When expectations are exceptionally high, “record but slightly below forecast” can be interpreted as the first sign that estimate momentum is slowing.

2. HBM4 shipments arrived slowly than expected: SK Hynix began mass shipments of HBM4 in the second quarter, but the main production ramp is scheduled for the second half. Management also said some high-value DRAM shipments moved into the second half, affecting the second-quarter product mix and average selling price.

DRAM bit shipments increased by a high-single-digit percentage sequentially and average selling prices rose approximately 30%. That is extraordinary in isolation. However, analysts had expected SK Hynix to capture more of the sharp increase in conventional DRAM prices. Official earnings deck

This creates an unusual short-term problem. SK Hynix’s strength is its high exposure to HBM, but conventional server DRAM prices rose so rapidly that the company’s HBM-heavy mix diluted its blended price increase.

3. Investors are questioning the rate of AI spending growth: The debate is not whether AI infrastructure spending will increase in 2026. It almost certainly will. The question is how quickly it grows after this year. UBS estimates that hyperscaler capital expenditure will rise 76% to $673 billion in 2026, followed by growth of 25% in 2027 and just 6% in 2028. Absolute spending would still rise, but semiconductor valuations often depend on expectations being revised upward, not merely remaining positive.

A separate Reuters analysis found that five major technology companies could add roughly $534 billion of capital spending between 2025 and 2027 while generating only about $340 billion of additional operating cash flow. That growing financing gap has made investors more sensitive to AI returns.

4. Capital expenditure is rising: SK Hynix expects 2026 capital expenditure in the high-₩40 trillion range, compared with ₩30.2 trillion in 2025. It is accelerating M15X, preparing the first Yongin facility and expanding advanced packaging capacity. Management says investment will be phased against confirmed demand and long-term customer agreements. That is sensible, but investors have seen this movie before. Memory companies invest during shortages, new supply arrives later, and profitability falls.

5. Investors wanted a clearer capital-return plan: SK Hynix ended June with ₩88 trillion of gross cash, ₩18.6 trillion of debt and ₩69.4 trillion of net cash. It then completed a $26.5 billion ADR offering after the quarter ended. At the July 28 exchange rate of ₩1,453.72 per dollar, the gross proceeds equal roughly ₩38.5 trillion. 

Adding this mechanically to June net cash produces approximately ₩108 trillion before fees, investment and other outflows. This is not a dividend pot because the proceeds are intended for manufacturing expansion. Still, it explains why investors expected greater clarity on shareholder returns. Management said it would communicate an updated plan later in 2026.

6. Leverage amplified the fall: The July 29 decline was not a clean referendum on SK Hynix’s fundamentals. The KOSPI fell 6%, leveraged single-stock products were being unwound, and margin pressure forced investors to reduce positions. SK Hynix fell nearly 20% intraday before closing 9.6% lower.

SK Hynix Q2 2026 Earnings: Revenue, Profit, Margins and Cash Flow

The operating results remained exceptionally strong.

Source: SK Hynix Q2 release and official earnings deck

The key accounting issue is net profit:

SK Hynix reported ₩122.7 trillion of pretax profit, including ₩63.27 trillion of investment-asset gains. These gains represented 51.6% of pretax profit. Analysts attributed much of the gain to the completed sale of the company’s Kioxia investment, although SK Hynix did not officially identify the asset in its results package.

Subtracting the identified gain leaves approximately ₩59.4 trillion of pretax profit. If we mechanically apply the quarter’s 23.46% effective tax rate, estimated net profit falls to roughly ₩45.5 trillion, with diluted EPS near ₩63,900 rather than the reported ₩131,478.

That is not company-adjusted EPS. Tax treatment could differ and other non-operating items remain. It simply shows why the 118% headline net margin should not be annualised.

The operating outlook, meanwhile, did not collapse. Management expects third-quarter DRAM bit shipments to rise approximately 10% sequentially, with NAND growing by a low-single-digit percentage. HBM4 is expected to ramp during the second half, while HBM4E volume production is targeted for 2027. The company has also concluded long-term agreements with around ten customers. 

Is SK Hynix Stock a Buy After the Q2 Earnings Drop? The 3P Test

A useful way to assess the correction in SK Hynix stock (SKHY) is the 3P Test.

1. Performance: Is SK Hynix’s Business Momentum Still Intact?

Not yet. SK Hynix’s revenue, operating profit, margins and cash generation all reached exceptional levels. HBM4 shipments have begun, enterprise SSD revenue doubled sequentially, and Solidigm’s revenue from high-capacity enterprise SSDs more than tripled.

The second-quarter miss appears partly connected to shipment timing and product mix, not disappearing demand. The test will come in the second half. If HBM4 ramps and DRAM pricing improves, the miss will look like delayed revenue. If HBM4 shipments slip again or SK Hynix continues to lag peer pricing, the explanation becomes less convincing.

2. Peak: Are SK Hynix’s Record Profits and Margins Sustainable?

This is the real debate. SK Hynix’s operating margin moved from 52% in 2018 to 10% in 2019, then to negative 24% in 2023. It reached 76% in Q2 2026. Memory earnings can reverse remarkably quickly.

HBM may make the next downturn different. Products are customised, qualification periods are lengthy and long-term agreements can include deposits, committed volumes and customer-specific pricing formulas.

But these agreements create a trade-off. During a rapid price increase, they can limit the supplier’s immediate upside. During a downturn, they may stabilise volumes. Their full protective value has never been tested through an HBM-led memory downturn.

Capacity is also returning. TrendForce estimates HBM will consume 22% of industry DRAM wafer input in 2026 and 30% in 2027, despite representing a much smaller percentage of memory bit output. That currently constrains conventional DRAM supply. If HBM growth slows after new fabs arrive, some capacity can eventually return to conventional products.

3. Premium: SKHY ADR vs SK Hynix Shares

This matters because each SKHY depositary share represents one-tenth of a Korean common share. On July 28, before the earnings release:

• Seoul-listed SK Hynix closed at ₩1.55 million.
• At ₩1,453.72 per dollar, the economic value per ADR was approximately $106.62.
• SKHY closed at $130.17.
• The resulting ADR premium was approximately 22.1%.

Note that SKHY’s July 28 close was pre-earnings, while the July 29 Seoul price was post-earnings. Comparing those two directly would produce a misleading 35% spread, so it should not be treated as an executable premium. The premium can persist because the ability to create new ADRs from Korean shares is constrained by the 17.79-million-share conversion cap. This limits the normal arbitrage mechanism that would pull the two prices together.

Think of it like buying gold jewellery in India. The underlying gold has a market price, but the buyer also pays a making charge. SKHY provides dollar trading, U.S. market hours and easier access, but investors may pay an additional “access charge.” That charge can shrink even when the underlying business performs well. Therefore, the fall in SKHY does not automatically mean it offers the same valuation as the corrected Seoul share.

SK Hynix Stock Price Targets and Analyst Ratings

Analyst opinions were unusually divided even before the results.

Sources: Barclays coverage, BNK and Korea Investment comparison, Korean brokerage targets. These targets predate the Q2 release and may be revised. 

The target-price difference is mainly a disagreement about 2027 and 2028, not Q2 2026. BNK forecasts net profit falling from ₩156.8 trillion in 2027 to ₩48.9 trillion in 2028. Korea Investment expects net profit to rise from ₩294.3 trillion to ₩351.3 trillion over the same period. A simple scenario model shows why both sides can call the stock differently.

The lesson is simple. SK Hynix looks extremely inexpensive if HBM turns memory into a more durable growth industry. It can still be expensive on depressed-cycle earnings if profits revert towards the BNK 2028 scenario. SKHY investors must then add the ADR premium. A 22% wrapper premium produces an effective valuation approximately 22% above the local share for the same earnings claim.

Key Risks and Catalysts for SK Hynix Stock in 2026 and 2027

NAND should also be watched as a possible early-cycle indicator. Enterprise SSD demand is currently strong, but TrendForce expects the 2026 NAND shortage to move towards balance during the second half of 2027 as process improvements and Chinese supply expand.

SK Hynix Stock Outlook: Better Entry or Peak-Cycle Trap?

SK Hynix’s business thesis remains largely intact. The company continues to generate extraordinary operating profit and cash, HBM4 shipments have started, and long-term customer agreements provide greater visibility than traditional commodity-memory orders. The stock thesis, however, is more complicated.

The correction has made the Seoul-listed share materially less demanding. But a 76% operating margin, rising capacity, slowing AI-capex growth and the memory industry’s history warn against treating peak earnings as permanent. For investors accessing SK Hynix through SKHY, the ADR premium is an additional risk. The underlying company may become cheaper while the U.S. wrapper still carries an access premium.

• SK Hynix stock bull case requires a successful HBM4 ramp, durable AI spending, disciplined capacity growth and evidence that long-term agreements reduce cyclicality. 
• SK Hynix bear case requires only a more familiar outcome: demand growth slows as new capacity arrives, margins normalise and investors stop paying for peak earnings.

The correction has improved the setup, but it has not removed the debate. Investors should track five things before drawing a conclusion: HBM4 execution, blended DRAM pricing, 2027 hyperscaler capex, capital allocation and the SKHY ADR premium. That checklist is more useful than trying to predict the exact bottom.