Snapshot: Executive Summary
SK hynix is the world's largest maker of high-bandwidth memory (HBM) and the #2 DRAM maker, headquartered in Icheon, South Korea. HBM is the vertically stacked DRAM bolted beside every AI GPU, and SK hynix is the clear leader: the lead NVIDIA HBM supplier, first to HBM4 mass production, and holder of a majority of NVIDIA's Rubin-generation HBM4 allocation4. On the library's "own the toll, not the traffic" frame, HBM leadership is as close to a toll-road as memory gets: NVIDIA's accelerators cannot ship without it, and SK hynix has been the first through NVIDIA validation each generation.
The problem is the price of that leadership. On book value, the durable anchor for a deeply cyclical business, SK hynix trades at ~10x current book (book value per share ₩231,952), an extreme against a ten-year 0.75x-2.9x band3. The spot P/E is distorted both ways (trailing ~39x on FY2025 EPS as earnings lag the cycle, forward ~7.5x on a FY26E EPS jump to ₩315,736), which is exactly why memory is valued through the cycle, not on a snapshot multiple. And the tape has cooled: after a near-vertical run to a ₩2,987,000 all-time high on June 25, the stock has pulled back ~20% to ₩2,378,5005, and it fell 2.9% on June 29 when Seoul unveiled a nearly $600bn Samsung / SK hynix capacity plan that revived memory-oversupply concerns for the out-years9. The HBM revenue-share lead is also normalizing, from ~70% a year ago to ~57% at end-2025, as Samsung and Micron qualify at NVIDIA (Counterpoint)4.
FY2025 was a record: revenue ₩97.15tn (+47% YoY), operating margin 48.55%, net margin 44.18% (net ~₩42.9tn), EPS ₩60,378, ROE 44.15%2; Q1 2026 printed record margins on revenue of ₩52.58tn: operating profit reached ₩37.6tn (a ~72% operating margin, ahead of the ~₩35.7tn consensus)8, net income rose +322% YoY, and net cash inflected to ~+₩32.5tn (cash ₩54.36tn vs debt ₩21.83tn)1. We initiate at Hold, 12-month PT ₩2,700,000 (bull ₩3,800,000 / bear ₩1,450,000) via a through-cycle EPS x normalized-multiple model. That ₩2.7M sits below the Street's median (₩3,300,000) and mean (₩3,334,250) against a Buy consensus (34 Buy + 4 Overweight / 2 Hold / 1 Sell)6: a deliberate valuation-discipline stance. The franchise is Buy-quality; the price, on book, is not here. Note the library's memory peer Micron (MU) is rated Buy at similar forward multiples, and SK hynix is not obviously cheaper.
Tactical: SK hynix is trading at ₩2,378,500, ~13% upside to our ₩2,700,000 12-month target (itself below the ~₩3.3M Street median). We rate the HBM leader a Hold: the clear #1 in the memory that matters, with 2026 sold out and net cash, but valued at ~10x current book with the share lead normalizing and out-year oversupply risk building as Korea funds a nearly $600bn capacity expansion. Memory is a through-cycle book-value and mid-cycle-earnings story, not a spot-multiple one; the risk/reward is balanced-to-asymmetric-down after a ~10x run. Rating Hold.
Investment Thesis
Bull Case
- The AI memory supercycle holds through 2027: HBM4 ramps on schedule at a majority NVIDIA Rubin allocation, and 2027 capacity fills as fast as 2026
- Book value catches up to the earnings surge, so the ~10x current-book multiple deflates to a defensible level as equity compounds
- FY26E EPS ~₩315,736 at ~12x → ₩3.79M, between the Street mean (₩3.33M) and the ₩5.3M high
- The HBM share lead stabilizes near ~57% rather than compressing toward Samsung and Micron
Base Case
- FY26E EPS ₩315,736 (WSJ consensus) at ~8.55x → ₩2.70M, cross-checked at ~11.6x current book (~5x the fast-rising forward book)
- The supercycle is real but the share lead normalizes and a capex-driven out-year oversupply risk caps the near-term multiple
- 2026 is sold out; net cash of ~₩32.5tn de-risks the balance sheet, but the equity story is peak-cycle
- A Buy-quality franchise where the book multiple already prices the win: we haircut the Street's optimism
Bear Case
- A post-2026 HBM correction as capacity expands, plus an NVIDIA Rubin cut, rolls over the record margins fast (peak-cycle operating leverage)
- The valuation resets from ~10x book toward the historical ~4-5x band as the cycle turns
- Value on a normalized mid-cycle EPS ~₩145,000 x ~10x → ₩1.45M, near the ₩1.4M Street low5
- NAND stays weak and the share lead keeps compressing as Samsung and Micron take Rubin allocation
Rating: Hold. The probability-weighted scenarios (50% base / 25% bull / 25% bear) blend to ~₩2.66M, just below the ₩2.7M base and above the ₩2,378,500 spot: a deliberate downside skew. We set the headline at the ₩2.7M base, which sits below the Street's median target (₩3,300,000). Note this is a more cautious view than a Buy sell-side consensus (34 Buy + 4 Overweight / 2 Hold / 1 Sell, mean ₩3,334,250): we will not fight the indispensable HBM leader with a Sell, but at ~10x current book with the share lead normalizing and out-year oversupply risk building, the risk/reward is balanced-to-asymmetric-down here. A Buy is defensible if you weight HBM leadership and sold-out 2026 over the near-term supply-cycle timing (Micron, at similar multiples, is a Buy in this library). Upgrade trigger: a pullback toward ~₩2.0M, confirmation the upcycle extends toward 2028 (as management and some analysts argue), or evidence the share lead stabilizes.
Business Overview
SK hynix makes memory chips: the volatile, high-volume commodity that stores and feeds data across every computer and phone, and now every AI accelerator. Its core is DRAM, where it is the #2 global maker and, critically, the #1 in HBM, the high-bandwidth stacked DRAM that the AI build-out is pulling forward. A weaker, more commoditized NAND flash leg rounds out the model, supported more by price hikes and supply discipline than by volume. The whole franchise is deeply cyclical: memory prices boom and bust, and SK hynix's earnings swing with them. The franchise's standing has re-rated with the cycle: in mid-2026 SK hynix overtook Samsung to become South Korea's most valuable company, both memory makers topped $1tn in market value, and beyond NVIDIA the company was selected by Microsoft to supply memory for its custom AI chips. It has also filed to raise ~₩45.5tn ($29bn) in a US ADR listing on Nasdaq7.
FY2025 revenue mix by product (est.)
DRAM (HBM-led) is the profit core; HBM exceeded ~40% of DRAM revenue and keeps rising; NAND is the weaker leg. The DRAM/HBM/NAND split shown is an author estimate against the WSJ revenue base.
Analyst ratings (WSJ/FactSet)
The sell-side is net Buy: 34 Buy + 4 Overweight vs 2 Hold and 1 Sell, a more bullish stance than our Hold, which sits below the Street median target.
Where SK hynix sits in the memory chain
- DRAM (the profit core): dynamic memory for servers, PCs, and phones. SK hynix is the #2 maker globally, and the operating profit runs almost entirely through DRAM, now increasingly HBM.
- HBM (the growth engine): high-bandwidth memory stacked beside AI GPUs. SK hynix is the clear #1 (~57% revenue share), the lead NVIDIA HBM supplier, and first to HBM4 mass production. HBM has exceeded ~40% of DRAM revenue and is still rising.
- NAND (the weaker leg): flash storage for SSDs and mobile. More commoditized and lower-margin; supported by price hikes and supply cuts rather than volume growth.
- The cycle: memory is a boom/bust capex business. FY2025 and 2026 are a supercycle peak; the durable through-cycle anchors are book value and normalized mid-cycle earnings.
HBM: The AI Memory Moat
The single most important section in this report. AI accelerators are gated not by compute alone but by memory bandwidth: how fast data reaches the GPU. HBM solves this by stacking DRAM dies vertically (16-high on HBM4) with a wide interface, delivering far more bandwidth per watt than standard DRAM. Every NVIDIA, AMD, and custom-silicon accelerator ships with HBM bolted alongside, and SK hynix is the supplier that has been first through NVIDIA's validation each generation. It moved to HBM4 mass production ahead of Samsung and Micron and holds a majority of NVIDIA's Rubin-generation HBM4 allocation4.
Revenue trajectory (₩tn)
FY23 (the memory trough) to FY25 ₩97.15tn actual; FY26E and FY27E are consensus: a steep ramp the HBM mix is meant to drive as HBM4 scales into 2027.
HBM revenue share (%)
SK hynix ~57% (Counterpoint, end-2025); Samsung ~22% / Micron ~20% author est. SK hynix is the clear #1, but the lead has normalized from ~70% a year ago as both rivals qualify at NVIDIA.
Why SK hynix, and why the moat is real
HBM is hard: stacking 16 DRAM dies with through-silicon vias, keeping thermals and yields in line, and clearing NVIDIA's exacting validation is a multi-year capability. SK hynix has been the first mover every generation (HBM2E, HBM3, HBM3E, now HBM4), which earns it the majority allocation and the pricing power. That is the "route through, not past" test from the library's synthesis: the AI build cannot avoid HBM, and it cannot avoid SK hynix as the lead supplier today. The record FY2025 margins (48.55% operating) are the moat cashing out. Management frames the current run as a "structural shift" rather than a normal cycle: with customers prioritizing procurement over price and HBM demand exceeding capacity over the next three years, it argues the upcycle can extend into 2027 and possibly 20288. Independent sell-side voices echo the framing: Goldman Sachs' Alvin So says this is "not the typical memory cycle" and sees "a structural change" in the industry11.
The wrinkle: the lead is normalizing, and the cycle can still turn
Two things temper the bull narrative. First, the revenue-share lead is compressing: from ~70% a year ago to ~57% at end-2025, as Samsung and Micron finally qualify HBM3E/HBM4 at NVIDIA (Counterpoint)4. A duopoly-plus-one erodes the pricing power a near-monopoly held. Second, and more structural, the memory cycle can turn: Korea's nearly $600bn Samsung / SK hynix capacity plan is set to double national DRAM capacity within five years, and Morningstar cautions that such build-outs "often lead to oversupply in tail years" if AI spending slows9. The franchise quality is not in question; the durability of the peak is, and a ~10x-book multiple leaves no room for it.
The earnings backdrop
Annual EPS trajectory (₩)2
From a FY2023 loss through the trough, to FY2025 EPS ₩60,378, to a FY26E consensus of ₩315,736 (a ~5x jump as HBM scales). The acceleration is real; the question the ~10x book multiple poses is how much of it is a durable peak.
DRAM, HBM & NAND: The Segments
Outside the HBM headline, SK hynix runs a three-product memory business whose economics diverge sharply. DRAM (now HBM-led) carries almost all the profit; NAND is the weaker, more commoditized leg. The bear case is less about the HBM moat than about the cycle riding this base: memory prices boom and bust, and SK hynix's record margins are a peak, not a plateau.
DRAM and HBM: the profit core
- DRAM (HBM-led): the profit engine. HBM has exceeded ~40% of DRAM revenue and keeps rising, dragging blended DRAM margins to record levels as AI demand outruns supply. This is where the FY2025 48.55% operating margin comes from.
- This core is also the source of the peak-cycle risk: any HBM or DRAM price rollover reverses record margins fast, because operating leverage cuts both ways.
NAND: the weaker leg
- NAND flash: storage for SSDs and mobile, a more commoditized and lower-margin business. Supported through 2026 by price hikes and supply cuts rather than volume growth.
- NAND is the drag on the blend: a weak leg that dilutes the DRAM/HBM strength and offers less through-cycle protection.
The risk lens
The base case is that 2026 capacity (sold out across HBM, DRAM, and NAND) fills at strong prices while HBM4 ramps; new wafer capacity only adds volume in 2027. The downside is specific: an HBM or DRAM price rollover as capacity expands, arriving while the multiple still pays ~10x book. That single dynamic (peak margins reversing on a priced-for-perfection tape) is the difference between the base and bear cases.
Financial Health & Trends
Revenue & operating margin (FY23–FY27E)
Price / book history (x)
The trough-to-record arc
SK hynix's revenue collapsed in the 2023 memory downturn (a full-year operating loss), then re-accelerated violently on the AI/HBM cycle. FY2025 closed at record levels: revenue ₩97.15tn (+47% YoY), operating margin 48.55%, net margin 44.18% (net ~₩42.9tn), EPS ₩60,378, ROE 44.15%2. Q1 2026 extended the run: revenue ₩52.58tn with operating profit of ₩37.6tn (a ~72% operating margin, above the ~₩35.7tn consensus)8 and net income up +322% YoY. Book value per share reached ₩231,952 and the balance sheet inflected to net cash. Consensus models FY26E EPS at ₩315,736 (a ~5x jump), with a Q2 2026 EPS estimate of ₩68,650.
Annual EPS: FY23 loss → FY26E ₩315,736
EPS in ₩; the operating-margin line is the reported/estimated rate. The FY2023 loss, the FY2024-25 recovery, then the FY26E surge. FY26E is consensus; the jump is what makes forward P/E look cheap and trailing look expensive.
FY2025 print highlights
| Metric | FY2025 | YoY | Driver | Note |
|---|---|---|---|---|
| Total revenue | ₩97.15tn | +47% | AI/HBM demand | reported |
| Operating margin | 48.55% | record | HBM mix / pricing | reported |
| Net margin | 44.18% | record | operating leverage | reported |
| Net income | ~₩42.9tn | +large | margin expansion | reported |
| EPS | ₩60,378 | +large | net income | reported |
| Return on equity | 44.15% | — | peak-cycle | reported |
| FY26E EPS (consensus) | ₩315,736 | +~5x | HBM ramp | WSJ (9) |
Note: FY2025 annual and Q1 2026 figures are from the WSJ Financials set; FY26E/Q2 26E are consensus. Book value per share ₩231,952 and net cash ~+₩32.5tn (cash ₩54.36tn, debt ₩21.83tn) are as of Q1 2026. Values are ₩ (won).
Capital Allocation & Returns
SK hynix has inflected to a net-cash balance sheet: cash of ₩54.36tn against debt of ₩21.83tn, ~+₩32.5tn net, after years of cyclical net debt1. Roughly 50% of accumulated free cash flow is pledged to shareholder returns for 2025-2027, and the dividend was raised to ₩3,000/share. But the equity story here is not income: the capital return is real, yet you own SK hynix for the HBM leadership and the through-cycle book value, not the ~0.1% dividend yield at this price.
Net cash / (debt) trajectory (₩tn)
Years of cyclical net debt through the 2023 trough, then a sharp inflection to ~+₩32.5tn net cash as the AI cycle generated record free cash flow. The de-risking is genuine and recent.
Annual EPS trajectory (₩)
FY23 loss → FY25 ₩60,378 → FY26E ₩315,736: the same surge that funds the net-cash inflection and the dividend raise, and that the ~10x book multiple already discounts.
The reinvestment question
SK hynix's best investment is its own roadmap: HBM4, 16-Hi stacking, and new fabs, funded from record operating cash. Capex was ~₩29tn in FY2025 and is guided "significantly higher" for 2026 to fund HBM4 and new capacity, including an ~$8bn order for ASML EUV lithography systems. At ~10x current book, the value question is not capital allocation (the balance sheet is now a fortress); it is whether the peak-cycle earnings that fund it are durable. That is the work of the next two sections.
Valuation Overview
Memory cannot be valued on a spot P/E, because a deep cycle distorts it both ways: SK hynix's trailing P/E is ~39x on FY2025 earnings (~23x on a trailing-twelve-month basis) (earnings lag the cycle) and its forward P/E is ~7.5x (the market prices a ~5x FY26E EPS jump to ₩315,736). Neither is the anchor. Through the cycle, the durable anchors are price-to-book and normalized mid-cycle earnings. On book, SK hynix trades at ~10x current book (BVPS ₩231,952), a genuine extreme against a ten-year 0.75x-2.9x band; trackers show ~3.8x on the stale FY25 year-end book, the classic memory book-lag as equity compounds faster than the reported figure. The valuation question is a durability question: does the supercycle hold long enough for book value to grow into the price, or does the cycle turn first?
The bull's strongest valuation counter is that this rally has been earnings-driven, not multiple-driven: the Kospi trades at roughly 8.6x forward earnings (versus ~21.2x for the S&P 500), SK hynix lower still at ~7.5x forward, and some managers (Robeco) call Samsung and SK hynix "quite cheap" on an earnings basis even after the run. We take the point, but anchor on book: forward EPS is a peak-cycle number, and it is the ~10x book value, not the P/E, that the market has never paid before11.
Price / book history (x)
The 0.75x-2.9x historical band vs today's ~10x current-book extreme. Memory has never sustained a multiple like this; the bull case needs book value to catch up fast.
Analyst target distribution (₩M)
WSJ: Low ₩1.4M · Median ₩3.30M · Mean ₩3.33M · High ₩5.3M, against a ₩2,378,500 spot and our ₩2.7M PT. Our base sits below the Street; bull ₩3.8M and bear ₩1.45M sit inside the Street range.
Why the scenario model
A single multiple obscures the two things actually in play: what a normalized mid-cycle EPS is worth, and what multiple the market pays on it. We value FY26E / normalized EPS × a through-cycle P/E → implied PT, cross-check against price-to-book, run three scenarios, and weight them. The base (FY26E EPS ₩315,736 at ~8.55x) reconciles to ₩2,700,000, cross-checked at ~11.6x current book (~5x the fast-rising forward book); the bear uses a normalized mid-cycle EPS (~₩145,000). Note the library's memory peer Micron (MU) is a Buy at similar multiples; SK hynix is not obviously cheaper. That work (with an interactive tool) is the next section.
Scenario Model: EPS × Normalized Multiple
The model is the heart of the report. Each scenario values memory through the cycle: EPS × a normalized P/E = implied PT, cross-checked against price-to-book. Toggle Base / Bull / Bear to load the anchor inputs; drag the EPS and P/E sliders to see the implied PT update. The base (FY26E EPS ₩315,736 at ~8.55x) reconciles to ₩2,700,000: below the Street median (₩3.30M). The bear deliberately uses a lower normalized mid-cycle EPS (~₩145,000), not the peak FY26E figure, because a bear case is a turned cycle.
The two live levers are FY26E/normalized EPS (₩) and the through-cycle P/E multiple. The P/B cross-check reads current book value per share ₩231,952 (so ₩2.70M ≈ ~11.6x current book, ~5x the fast-rising forward book). Memory is a book-value and mid-cycle-earnings story, so both the EPS and the multiple are the swing variables.
| Line | Value |
|---|---|
| FY26E / normalized EPS | ₩315,736 |
| × through-cycle P/E | 8.55x |
| = implied PT | ₩2,700,000 |
| P/B cross-check (÷ ₩231,952 book) | — |
| vs current market | — |
EPS and P/E load from the scenario anchor and are user-editable via the sliders. Base reconciles to ₩2,700,000 (FY26E EPS ₩315,736 × 8.55x); bull ₩3,800,000 (₩315,736 × 12.0x); bear ₩1,450,000 (normalized EPS ₩145,000 × 10.0x, near the ₩1.4M Street low). The P/B line cross-checks the implied PT against current book value per share ₩231,952.
Sensitivity grid: EPS × P/E (PT in ₩)
PT calculator (EPS × multiple)
Risk / Reward calculator
Ask the Thesis AI-assisted checking…
Describe a scenario in natural language; the assistant returns a structured impact analysis against this dashboard's thesis, scenarios, and scenario-model math. Powered by Claude via a Cloudflare Worker proxy (Anthropic key held server-side; same pattern as the live-quote feed).
Note: The assistant reasons from the dashboard's data snapshot and thesis sections: it does not browse the web or access real-time fundamentals beyond what's in data.js. Treat its responses as scenario-modeling support, not primary research. Author judgments on rating, PT, and probabilities remain with the analyst.
Upcoming Catalysts
| Catalyst | Window | Why it matters |
|---|---|---|
| Q2 2026 earnings | Jul 29, 2026 | Q2 EPS estimated ~₩68,650; the read on whether the record ~72% operating margins hold and HBM/DRAM pricing stays firm is the key near-term event. |
| HBM4 volume ramp | Q4 2026 | HBM4 shipments start in Q4 2026 into NVIDIA Rubin: SK hynix holds the majority Rubin HBM4 allocation, and this ramp must land on schedule for the base and bull cases. |
| 16-Hi HBM4 mass production | Q3 2026 | Targeted Q3 2026: the next stacking milestone that extends the technology lead over Samsung and Micron if hit on time. |
| 2026 memory pricing | Each quarter | DRAM contract prices up strongly QoQ; sustained pricing power is what keeps the record margins from rolling over, and the bear case from engaging. |
| Capital-return path | 2025–27 | ~50% of accumulated FCF pledged to returns and the dividend raised to ₩3,000; execution on net cash and buybacks supports the through-cycle case. |
| US ADR / Nasdaq listing | ~Jul 10, 2026 | A ~₩45.5tn ($29bn) Nasdaq ADR listing (~2.5% of shares outstanding) is expected to begin trading around July 10, funding new domestic fabs; it broadens the shareholder base and could narrow the valuation gap to Micron. A structural, not fundamental, catalyst.7 |
Risk Factors
- Peak-cycle operating leverage. The pre-eminent risk. Record FY2025 margins (48.55% operating) are a cyclical peak; any HBM or DRAM price rollover reverses them fast, because leverage cuts both ways. Memory has never sustained margins like these, and the whole bear case is this peak turning.
- NVIDIA concentration. HBM revenue leans heavily on NVIDIA, the primary customer for SK hynix's Rubin-generation HBM4. A cut to NVIDIA's build plans, or a shift of allocation to Samsung and Micron as they qualify, would hit the demand the whole thesis depends on.
- The HBM share lead is normalizing. From ~70% a year ago to ~57% at end-2025 (Counterpoint), as Samsung and Micron qualify HBM3E/HBM4 at NVIDIA. A duopoly-plus-one erodes the pricing power a near-monopoly held.
- Capacity-driven oversupply. Korea's nearly $600bn Samsung / SK hynix plan is set to double national DRAM capacity within five years; Morningstar warns such build-outs "often lead to oversupply in tail years" if AI spending slows, since fresh fabs take two to three years to come online and can arrive as demand tapers9.
- Memory-efficiency software. Techniques such as Google's TurboQuant that reduce the memory an AI model needs could, at the margin, soften demand growth; SK hynix argues they instead expand the market by improving AI economics.
- Valuation / multiple compression. At ~10x current book (against a 0.75x-2.9x band), the multiple is priced for perfection. Any pricing, demand, or timing disappointment compresses it toward the historical ~4-5x, which is most of the bear-case move.
- Market frothiness and retail leverage. SK hynix and Samsung together now make up more than half of the Kospi, and the Korea Exchange delayed the June 2026 launch of single-stock weekly options (SK hynix among them) after record volatility, including an ~8% single-day drop followed by an ~8% rise the next day and a market volatility gauge at an all-time high. Regulators voiced "regret" over leveraged single-stock ETFs tied to the two chipmakers, and national margin debt sits at a record ₩37tn ($24bn) with about half of adults now holding brokerage accounts. A momentum- and leverage-heavy, retail-driven tape can amplify a drawdown if the AI narrative cools, independent of the fundamentals10.
- China / geopolitical exposure. SK hynix runs fabs at Wuxi and Dalian in China on annual US export licenses; a licensing or policy change could disrupt a meaningful part of the footprint.
- NAND weakness. The NAND leg is commoditized and lower-margin, supported by price hikes and supply cuts rather than volume; it offers less through-cycle protection than the DRAM/HBM core.
- Cyclicality of the memory base. Memory is a boom/bust capex business; the 2023 downturn (a full-year operating loss) is a recent reminder that the trough can be severe even with the HBM moat intact.
Scenario Stress Tests
| Scenario | Mechanism | Anchor PT | Delta vs base ₩2.70M |
|---|---|---|---|
| Base | FY26E EPS ₩315,736 × 8.55x (≈11.6x current book) | ₩2,700,000 | — |
| Multiple resets toward the band | ₩315,736 EPS × 7.0x | ~₩2,210,000 | (18%) |
| Rubin trim: EPS haircut | ₩250,000 EPS × 8.55x | ~₩2,138,000 | (21%) |
| Cycle turns, book-multiple compression | Normalized EPS ₩180,000 × 9.0x | ~₩1,620,000 | (40%) |
| Full bear (HBM correction + Rubin cut) | Normalized EPS ₩145,000 × 10.0x | ₩1,450,000 | (46%) |
| Bull: supercycle holds | ₩315,736 EPS × 12.0x | ₩3,800,000 | +41% |
All stress-test PTs are derived from the same EPS × normalized-multiple framework used in the interactive tool: each shock changes only the inputs noted. Deltas are vs the ₩2,700,000 base PT. The bear cases deliberately use a normalized mid-cycle EPS (not the peak FY26E figure), because a bear case is a turned cycle. Several adverse scenarios sit between the ₩2,378,500 spot and the bear anchor, which is why the probability-weighted blend (~₩2.66M) is just below the base and only modestly above spot.
Bull vs Bear Debate
| Issue | Bull view | Bear view |
|---|---|---|
| Is the moat real? | SK hynix is the clear #1 in HBM, the lead NVIDIA HBM supplier, first to HBM4 mass production, holding a majority of Rubin allocation. 2026 is sold out and it printed a record FY2025. A genuine chokepoint. | Moats protect share, not the multiple. Samsung and Micron are qualifying at NVIDIA; the ~57% share (from ~70%) is normalizing, and ~10x book can compress hard even with the lead intact. |
| Cycle durability | The AI memory supercycle is structural: HBM demand outruns supply, 2026 is sold out, and new wafer capacity only adds volume in 2027. The peak can persist. | Memory always mean-reverts. Record 48.55% operating margins are a peak, not a plateau; peak-cycle leverage reverses fast on any price rollover. You are paying a cyclical top. |
| NVIDIA / Rubin | SK hynix holds the majority Rubin HBM4 allocation and cleared validation first; any near-term demand jitters are noise against a multi-year, sold-out build. | Heavy NVIDIA concentration means a single customer's build decision can dent the whole demand base, and Samsung and Micron are qualifying to take share of the Rubin allocation. |
| Valuation | On forward EPS (~7.5x) it is not expensive, and net cash of ~₩32.5tn plus a ₩3.3M Street median support a Buy. Micron, at similar multiples, is a Buy in this library. | Spot P/E is a distorted memory trap. On book (~10x, a historic extreme) it is priced for perfection, and it is not obviously cheaper than Micron on the durable anchors. |
| Our call | A Buy is defensible on HBM leadership, sold-out 2026, and net cash. | We Hold at ₩2.7M (below the median): Buy-quality franchise, but the book multiple already prices the win, and out-year oversupply risk is building. |
Technical Analysis
SK hynix trailing-12-month closes (₩)
RSI (multi-timeframe)
Cooled hard off the June ₩2,987,000 all-time high, consistent with a stock unwinding a parabolic move.
MACD vs Signal
Constructive through the AI/HBM run but rolling over as the parabolic move unwound near the highs.
Relative strength (2026 YTD)
SK hynix crushed the SOX and the KOSPI year-to-date (+285.5%) on the HBM supercycle: a crowded, momentum-heavy tape now cooling.
EMA stack (current, ₩)
Trader's view
- Price > 50-DMA > 200-DMA: still a bullish stack after the ~12x move, but sharply extended above the longer averages and now pulling back.
- Key support: the ₩2,000,000 area (a round-number shelf below spot). A close below would signal the HBM trade is unwinding faster.
- Key resistance: the ₩2,987,000 all-time high. A decisive break needs a fresh catalyst (a blowout Q2, or confirmation the upcycle extends into 2028).
- Momentum: rolling over off the highs after a parabolic run, consistent with a Hold that trades just below its ₩2.7M target after a ~20% pullback from the ATH. The pullback also tracks a market-wide Kospi volatility spike that tripped an unprecedented number of trading curbs and pushed the index's volatility gauge to a record10.
Glossary & Methodology Notes
- HBM (High-Bandwidth Memory)
- Vertically stacked DRAM (16-high on HBM4) with a wide interface, delivering far more bandwidth per watt than standard DRAM. Bolted beside every AI GPU. SK hynix is the clear #1 (~57% revenue share) and the lead NVIDIA HBM supplier.
- DRAM vs NAND
- DRAM is fast, volatile working memory (the profit core, now HBM-led); NAND is slower, non-volatile flash storage (the weaker, more commoditized leg). SK hynix is #2 in DRAM globally and #1 in HBM.
- The memory cycle
- Memory is a boom/bust capex business: prices and margins swing violently with supply and demand. FY2025-26 is a supercycle peak; the 2023 downturn was a full-year operating loss. This is why spot multiples mislead.
- Price-to-book vs P/E for cyclicals
- For a deep cyclical, a spot P/E is distorted (trailing looks expensive as earnings lag; forward looks cheap as the market prices the jump). Price-to-book and normalized mid-cycle earnings are the durable anchors. SK hynix is at ~10x current book vs a 0.75x-2.9x band.
- HBM4 / 16-Hi stacking
- The next HBM generation, stacking 16 DRAM dies. SK hynix moved to HBM4 mass production ahead of Samsung and Micron; 16-Hi HBM4 mass production is targeted Q3 2026, with volume shipments into NVIDIA Rubin from Q4 2026.
- NVIDIA Rubin
- NVIDIA's next-generation accelerator platform (after Blackwell), which uses HBM4. SK hynix holds a majority of the Rubin HBM4 allocation and is NVIDIA's primary HBM supplier; a hypothetical Rubin build cut is the central demand risk in the bear case.
- "Memflation"
- Shorthand for the sharp memory-price inflation of the AI cycle: DRAM and HBM contract prices rising strongly QoQ as demand outruns supply. It drives the record margins and, if it rolls over, the bear case.
- Net cash
- Cash and equivalents minus total debt. SK hynix inflected to ~+₩32.5tn net cash (cash ₩54.36tn vs debt ₩21.83tn) in Q1 2026 after years of cyclical net debt: a genuine balance-sheet de-risking.
Methodology
- Snapshot anchor: July 2, 2026. Live price patches via the Cloudflare-Worker quote proxy (Yahoo 000660.KS, in KRW) on page load.
- KRW-native: all market data and operating financials are in Korean won (the KRX listing); no FX conversion is applied.
- FY26E/Q2 26E figures are consensus (WSJ); reported actuals are from SK hynix Q1 2026 and FY2025 results (WSJ Financials). Scenario inputs and segment estimates are the author's view, flagged where used.
- The scenario model values FY26E / normalized EPS × a through-cycle P/E, cross-checked against price-to-book. The base reconciles to ₩2,700,000; the bear uses a normalized mid-cycle EPS.
- Conclusions are the author's view. Illustrative, not investment advice.
Sources & Citations
Inline citations
Superscripted numbers in the body link here. Click any N in the report to jump back to the source.
- SK hynix Inc., Q1 2026 results (reported April 22, 2026): revenue ₩52.58tn, record operating and net margins, net income +322% YoY; cash ₩54.36tn vs debt ₩21.83tn (net cash ~+₩32.5tn); book value per share ₩231,952. Figures from WSJ Financials (000660.KS). WSJ Financials. ↩ ↩
- SK hynix Inc., FY2025 annual results: revenue ₩97.15tn (+47% YoY), operating margin 48.55%, net margin 44.18% (net ~₩42.9tn), EPS ₩60,378, ROE 44.15%. FY26E EPS ₩315,736 (WSJ consensus, 9 analysts); Q2 2026 EPS estimate ₩68,650. Figures from WSJ Financials (000660.KS). WSJ Financials. ↩ ↩ ↩
- Valuation anchors: price ₩2,378,500 (WSJ, Jul 2 2026), ~728M shares, market cap ~₩1,732T; trailing P/E ~39x on FY2025 EPS (~23x TTM), forward P/E ~7.5x (on FY26E EPS ₩315,736), price-to-book ~10x on current book value per share ₩231,952 (~3.8x on the stale FY25 year-end book). The 0.75x-2.9x historical price-to-book band is a ten-year range (trade-press/tracker estimate, not a filing). ↩
- HBM market position: SK hynix ~57% HBM revenue share as of end-2025, ahead of Samsung and Micron (Counterpoint Research, reported by the Financial Times); the lead has normalized from an estimated ~70% a year earlier as both rivals qualify at NVIDIA. SK hynix moved to HBM4 mass production ahead of Samsung and Micron and holds a majority of NVIDIA's Rubin-generation HBM4 allocation (TrendForce estimate). Share and allocation figures are trade-press estimates, not a SK hynix filing. ↩ ↩ ↩ ↩
- Price action: SK hynix reached a ₩2,987,000 all-time high on June 25, 2026 and has since pulled back ~20% to ₩2,378,500 (WSJ market data, 000660.KS), unwinding a near-vertical AI/HBM run. The bear case values a normalized mid-cycle EPS ~₩145,000, near the ₩1.4M Street low. ↩ ↩
- Consensus (WSJ, Jul 2026): analyst ratings 34 Buy + 4 Overweight / 2 Hold / 1 Sell → consensus Buy; 12-month price target mean ₩3,334,250 / median ₩3,300,000 (range ₩1.4M–₩5.3M). Market data (price ₩2,378,500, ~728M shares, market cap ₩1,732T, 52-week range ₩245,000–₩2,987,000, forward P/E ~7.5x) from WSJ (000660.KS) as of July 2, 2026. ↩
- US listing (Financial Times, reported June 24, 2026): SK hynix filed to raise ~₩45.5tn ($29bn) via American depositary receipts on Nasdaq, expected to begin trading around July 10, 2026; the shares initially offered represent ~2.5% of outstanding stock, with proceeds funding new domestic chip plants. Analysts noted a US listing could narrow SK hynix's valuation gap to Micron. SK hynix is the primary HBM supplier to NVIDIA and was selected by Microsoft to supply memory for its custom AI chips; in mid-2026 it overtook Samsung as South Korea's most valuable company. ↩ ↩
- Q1 2026 results and "structural shift" (Financial Times / Bloomberg, reported April 22, 2026): operating profit ₩37.6tn ($25.4bn) on revenue ₩52.6tn, ahead of the ₩35.7tn consensus operating profit (~72% operating margin); revenue nearly tripled and earnings rose roughly fivefold year on year. Management described a "structural shift," with customers prioritizing procurement over price and HBM demand exceeding capacity over the next three years; one cited academic expected the upcycle to extend through 2027 and possibly 2028. SK hynix also announced ~$8bn of ASML EUV lithography purchases. ↩ ↩ ↩
- Korean capacity plan and oversupply risk (Financial Times, reported June 29, 2026): Samsung and SK hynix will invest a combined ₩911tn ($590bn) alongside the government in a nearly $600bn build-out; Seoul expects national DRAM capacity to double within five years. SK hynix closed down 2.9% on the announcement. Morningstar's Jing Jie Yu cautioned that such build-outs "often lead to oversupply in tail years" if AI spending slows, as fresh capacity takes two to three years to come online. ↩ ↩ ↩
- Market frothiness and single-stock options delay (Financial Times, reported June 25, 2026): the Korea Exchange postponed the June 29 launch of weekly single-stock options tied to Samsung, SK hynix, Hyundai Motor and LG Energy Solution, citing record volatility. The market saw an ~8% single-day drop followed by an ~8% rise the next day; the Kospi itself plunged ~10% in one session before recovering ~9%, its wild swings triggered an unprecedented number of trading curbs, and its volatility gauge hit a record high of 95. Samsung and SK hynix together now make up more than half of the Kospi, and the exchange's chief executive warned the concentration could exacerbate volatility. South Korea's Financial Supervisory Service expressed "regret" over leveraged single-stock ETFs tied to the two chipmakers; national margin debt stands at a record ₩37tn ($24bn), with about half of adults now holding brokerage accounts (up from 21% in 2019). ↩ ↩
- Korean market valuation and "structural change" (Financial Times, reported May 19, 2026): the Kospi more than tripled in under 18 months (2,401 to 7,272 points), six months faster than the Nasdaq Composite into its 2000 peak, yet trades at ~8.6x forward earnings versus ~21.2x for the S&P 500 (Bloomberg). JPMorgan Private Bank's Cameron Chui characterized the rally as "earnings driven" rather than "multiple driven"; Robeco's Joshua Crabb called Samsung and SK hynix "quite cheap" on an earnings basis; Goldman Sachs' Alvin So said "this is not the typical memory cycle" and sees "a structural change." Samsung and SK hynix's combined market value (~$2tn) exceeds South Korea's GDP (~$1.93tn, IMF estimate). ↩ ↩
Background reading
- SK hynix quarterly & annual results (Q1 2026, FY2025): revenue, operating margin, net income, book value, net cash.
- SK hynix capital-return policy (~50% of accumulated FCF, 2025-2027) and the ₩3,000/share dividend.
- Counterpoint Research / TrendForce on HBM revenue share and NVIDIA HBM4 (Rubin) allocation.
- Coverage of the HBM4 volume ramp (Q4 2026 shipments) and 16-Hi HBM4 mass production (Q3 2026 target).
- Financial Times / Bloomberg (2026): the Q1 2026 "structural shift" results, the $29bn US ADR listing filing and Microsoft custom-AI-chip selection, the nearly $600bn Korean capacity plan, and the Kospi's AI-driven, retail-led rally and record-volatility backdrop (the delayed single-stock options).
- Memory peer disclosures (Micron, Samsung): for the HBM-share and cross-cycle comparison.
- Market data: WSJ (000660.KS), stockanalysis.com (price, shares, market cap, 52-week range, multiples) as of July 2, 2026.
Disclaimer. This report is the author's institutional equity-research view, prepared for portfolio and educational purposes. It is not a recommendation to buy, sell, or hold any security. Forward-looking statements are subject to risk and uncertainty; past performance is not indicative of future results. Consult a licensed financial advisor before making investment decisions. All third-party trademarks are the property of their respective owners.