The Invisible Monopoly, The Last Gateway to Advanced Processes (HPSP — The World's Only High-Pressure Hydrogen Annealing)

SEMICONDUCTOR EQUIPMENT · 2026.06.15

The Invisible Monopoly,

The Last Gateway to Advanced Processing

HPSP — The World's Only High-Pressure Hydrogen Annealing

403870 · KOSDAQ

Equipment Driven by HBF, 2nm, and New Fabs All at Once — And the Breakthrough Above Target Price

June 15, 2026 · Narrative constructed after Goguma Team research and multi-round discussions
Invisible Monopoly · Gateway to Advanced Processing · New Application: HBF · Bull & Ultra-Bull Take-Profit Levels

Onion (Macro/Theme) · Acorn (Industry/Reality) · Peanut (Risk/Value) · Carrot (Chart/Flow) · Sweet Potato (Comprehensive)

Prologue

The Day It Broke Through the Target Price, What the 2.7 Trillion Won Trading Volume Said

On June 15, 2026, a small-to-mid-cap semiconductor equipment stock soared +16.78% in a single day, recording a trading volume of 2.7 trillion won. This was a level that broke through the brokerage consensus target price (approx. 68,000 won) by over 22%. HPSP — a name unfamiliar to many, yet every cutting-edge semiconductor factory in the world cannot make chips without this company's equipment. It was the day an 'invisible monopoly' stepped into the market spotlight.

HPSP manufactures High-Pressure Hydrogen Annealing (HPA) equipment — as semiconductors shrink below 10 nanometers (nm), microscopic defects (dangling bonds/traps) form at the interface between the insulating layer and the channel; if these are not healed, the chip fails to function properly. HPSP fills these defects by force-injecting 25 atm of ultra-high pressure and 100% pure hydrogen at low temperatures below 450 degrees.Advanced processing is impossible without this equipment, which is why their operating profit margin exceeds 50%.

This report answers two questions. Why has this 'invisible monopoly' exploded now — the answer lies in the simultaneous convergence of three demands: HBF (next-gen memory), 2nm, and new fabs. And now that it has broken through the consensus, how high can the stock go — 110,000 won, or perhaps 150,000–200,000 won? Based on multiple team discussions and data, we map out the final take-profit levels.

Chapter 1

The Invisible Monopoly — What is High-Pressure Hydrogen Annealing

The essence of HPSP is 'one irreplaceable thing.' They are the only company in the world that makes high-pressure-certified hydrogen annealing equipment, and this is reflected in their financial proof of monopoly: a 50% operating profit margin.

As semiconductors shrink, 'oxygen vacancies' and defects form at the boundary between the transistor's insulating layer (High-K), metal gate, and channel. These defects increase leakage current, destabilize threshold voltage, and degrade reliability and data retention. In the past, high-temperature heat treatment was used to heal them, but advanced devices are heat-sensitive, making high temperatures unusable.HPSP's solution is pressure instead of temperature — by injecting 25 atm of ultra-high pressure and 100% pure hydrogen at temperatures below 450 degrees, they strongly heal defects even with low heat.

The roots of this technology lie in the high-pressure chamber technology acquired after the bankruptcy of U.S.-based Gasonics in 2005. After over 10 years of development, TSMC adopted it in the mid-2010s as the only solution for yield limits in sub-10nm processes. Since becoming an independent entity in 2017, they now supply all global advanced fabsincluding TSMC, Samsung Electronics, SK Hynix, and Intel. They are the only company in the world to have obtained high-pressure certification, and that monopoly is proven by their 50% operating profit margin.

50%+

Operating Profit Margin (Consistent 2023–2026E) — An abnormally high margin in the equipment industry, proof of monopoly power and pricing authority.

Chapter 2

Why Explode Now — Three Demands Converging at Once

HPSP's equipment demand is tied 1:1 to 'advanced process transitions.' And in 2026, three massive demands driving that transition are converging simultaneously — 2nm, HBM, and HBF.

① 2nm GAA — Interface Quality Determines Yield

In the 2nm process, where semiconductor architecture shifts from FinFET to Gate-All-Around (GAA), the gate surrounds the channel from all sides, making interface quality and variability control significantly more critical. With Samsung, Intel, Rapidus, and TSMC all pursuing 2nm, the demand for low-temperature high-pressure hydrogen passivation (HPA) to improve yields is increasing.

② HBM & New Fabs — Expansion Equals Equipment Demand

SK Hynix is accelerating investments in M15X and the Yongin cluster, ordering approximately 8 trillion won worth of ASML's Extreme Ultraviolet (EUV) lithography equipment. New fabs are being constructed one after another, from Samsung Pyeongtaek (P4) to Intel's U.S. plants and Micron. HPSP, a front-end equipment stock, is directly exposed to this new fab cycle — with each piece of equipment priced at 3–6 billion won, 5–15 units per new fab line, and 15–30 units for large multi-line investments, a single fab contributes 20–80 billion won in revenue to HPSPalone.

③ HBF — A New Application Covered in the Next Chapter

The newest engine is HBF. It is a next-generation memory that SK Hynix and SanDisk have begun standardizing, which could open up an entirely new demand source for HPSP: 'NAND front-end processing.' We will cover this story in detail in the next chapter.

Key points of Chapter 2. HPSP's surge is not a theme; it's 'process physics' — as miniaturization progresses, healing interface defects becomes the key to yield, and HPSP is the only equipment that solves this at low temperatures. 2026–2028, when four demands (2nm, HBM, new fabs, HBF) converge, is the harvest season for this monopoly.

Chapter 3

HBF — Where NAND Becomes Memory

The hottest fuel for today's surge is HBF (High Bandwidth Flash). And the reason that fuel is significant for HPSP is that HBF is a technology that 'uses NAND like memory.'

HBF is a next-generation memory for which SK Hynix and SanDisk have initiated global standardization, aiming for a new tier between High Bandwidth Memory (HBM) and standard storage (SSD). It provides 'high-bandwidth, high-capacity storage' that feeds massive models and data quickly to AI inference servers. Kioxia's prototypes are mentioned at the 5TB/64GB/s level, with a schedule of prototypes in H2 2026 and integration into AI accelerators by early 2027.

The key is that HBF has a structure where 3D NAND is stacked vertically (die-stacked) and connected via hybrid bonding. To use NAND more like memory, cell characteristic variations, traps, data retention, read interference, and above all, interface defect management become significantly more important. After bonding 300+ layer high-stack NAND using hybrid bonding, there is almost no margin for heat (thermal budget) — at this point, High-Pressure Hydrogen Annealing, which heals defects at low temperatures, becomes a new candidate for adoption._"HBF Mass Production = Increased HPA Adoption in NAND Front-End"_is the investment logic derived here.

Key points of Chapter 3 — The direction is correct, but it is not yet an 'order'. The direction of HBF opening a new application (NAND front-end) for HPSP is clear. However, at this point, "mass orders for HBF-use HPA equipment" have not been confirmed by public disclosure — because HBF itself is an early-stage technology heading toward 2027 mass production. In other words, today's stock price is in the realm of 'expectations,' pricing in 2027–2028 revenue in advance. The moment these expectations turn into 'order disclosures' is the trigger for the next revaluation.

Chapter 4

Numbers — 50% Operating Profit Margin, Proof of Monopoly

HPSP's fundamentals are simple and powerful. Revenue fluctuates according to the new fab cycle, but the 50% operating profit margin remains unshaken — that is the essence of a monopoly.

Indicators	2025	2026E	2027E(Base)
Revenue	173 billion won	237.3 billion won (+37%)	Approx. 300 billion won [est.]
Operating Profit	89.9 billion won	126.6 billion won	Approx. 156 billion won [est.]
Operating Profit Margin	52.0%	53.4%	Approx. 52%
EPS	870 won	1,405 won	Approx. 1,640 won [est.]

The consistency of the operating profit margin is key. 53.2% in 2023, 51.8% in 2024, 52.0% in 2025, and a projected 53.4% in 2026 — maintaining a 50% operating profit margin for years in the equipment industry means there is almost no competition and the company sets the price. 2026 revenue is expected to grow +37% year-over-year, marking the beginning of the reflection of 2nm, HBM, and new fab demand.

Key points of Chapter 4. HPSP is not an 'equipment stock with volatile revenue,' but an 'equipment stock with fixed monopoly margins.' As new fabs increase, revenue rises in steps, and half of that revenue drops to the bottom line. The question is whether this monopoly will be maintained during the harvest season of HBF, 2nm, and new fabs, even with competitors entering the market.

Chapter 5

Four Perspectives — 110,000 Won or 200,000 Won?

Opinions are divided on the stock price, which has broken through the consensus. The team discussed whether it will go further or if it is overheated. The range varied from the 50,000-won level (overheated) to 200,000 won (ultra-long-term monopoly premium).

Onion · Macro/Theme (Bullish)

"High-Pressure Hydrogen Annealing is 'process physics' tied 1:1 to advanced process transitions — it's not a theme. On top of 2nm GAA, HBM, and new fabs, if HBF adopts HPA in large quantities for NAND front-end, 2027–2028 earnings will level up. With a monopoly and 50% operating profit margin, a PER 50x premium is justified. 'Bull' is at current price levels; 'Ultra-Bull' opens when HBF figures are confirmed."

Acorn · Industry/Reality (Balanced)

"It's a great company, but it's already expensive. The 2026 projected PER of 59x and PBR of 23x are prices that have already priced in the HBF revaluation. If HBF enters actual mass production investment and supplies NAND-use HPA meaningfully, 2027–2028 profits will rise, but if it's just a rumor, it's overheated. To justify 83,500 won, we need at least 2027 EPS of 1,600 won and a PER of 50x."

Peanut · Risk/Value (Cautious)

"The real risk isn't valuation; it's patents. In the June patent ruling, HPSP won the invalidation trial, but YEST won the negative scope confirmation trial — meaning HPSP's defense by narrowing the scope of key patents opened a bypass. YEST has already confirmed supply to two global memory makers in 2026. The 95% market share monopoly is already cracking, and if it breaks, the PER of 50–80x will drop immediately."

Carrot · Chart/Flow (Practical)

"+46% in three days from 57,000 won on June 9, with 2.7 trillion won in trading volume at the daily limit. It's clearly in the overbought record-high zone, but if that volume was supported by foreign and institutional net buying, it could be the start of an eruption. The previous day's limit price of 71,500 won is a strong primary support. Don't chase; take partial profits at 92,000 and 110,000 won, and let the trend decide anything above that."

Although the four perspectives didn't converge into a single point, the framework is clear. HPSP is a 'true monopoly tied to process physics,' but the stock price has already priced in future demand significantly in advance. Therefore, further upside depends on two conditions: ① whether HBF turns from 'expectation' to 'order disclosure,' and ② whether the monopoly (90%+ share) is maintained despite YEST's entry. If both are met, the path opens beyond 110,000 won to 150,000–200,000 won; if either breaks, a retracement to the 50,000–60,000-won level awaits.

Chapter 6

Valuation — Bull & Ultra-Bull Take-Profit Levels

The target prices below are five paths calculated by multiplying earnings forecasts (EPS) by the monopoly premium (PER). The key is 'which year's earnings you are buying' — if you buy 2026, it's overheated; if you buy the full-scale 2028 HBF/multi-fab era, 200,000 won is arithmetically justifiable.

HPSP (403870) — Current approx. 83,500 won

Scenario	Target Price	Basis
Bear	56,000 won -33%	YEST monopoly crack/HBF expectation fades. 2026E EPS 1,405 won × PER 40x.
Base	90,000 won +8%	2nm/HBM/new fab demand. 2027E EPS 1,640 won × PER 50x (monopoly premium).
Bull	130,000 won +56%	Early HBF adoption + expanded new fab orders. 2027E EPS approx. 2,000 won × PER 55x.
Ultra-Bull	150,000–200,000 won +80~140%	2028 HBF NAND front-end mass adoption + multi-fab (Samsung/SK/Intel/Micron/Rapidus) simultaneous operation + 90%+ share maintenance. 2028E EPS 3,000–4,000 won × PER 50x.

The current price has already surpassed the consensus (approx. 68,000 won) and the Base — partial profit-taking is standard rather than chasing. The previous day's limit price of 71,500 won is primary support; breaking 64,000 won damages the trend. Take partial profits at 92,000 and 110,000 won; hold the Ultra-Bull area for the trend. The two triggers for further upside are 'HBF order disclosure' and 'market share maintenance'.

The path to 150,000–200,000 KRW. This 'Super Bull' case is not fiction, but arithmetic — if HBF is mass-adopted in the NAND front-end process by 2028 and multiple new fabs begin operations simultaneously, the EPS of this monopoly with 50% operating margins could reach 3,000–4,000 KRW. Multiplying this by a monopoly-level PER of 50x yields 150,000–200,000 KRW. However, this path is a ultra-long-term bet on 'buying 2028's earnings today,' and it must pass three checkpoints: ① actual progress in HBF mass production, ② simultaneous operation of new fabs, and ③ maintaining a monopoly despite YEST's entry. If any of these are blocked, the 110,000 KRW range becomes the ceiling.

Chapter Seven

Signs of the Monopoly Cracking

The biggest premise for this stock price is the 'continuation of the monopoly.' Therefore, we define in advance the signals that could shake this monopoly and dampen expectations.

Signal	Meaning	Action
① YEST expands market share	If YEST moves beyond two memory clients into foundry and advanced processes, the 95% monopoly effectively breaks.	Immediate de-rating of PER to 50–80x. Revaluation of the monopoly premium.
② Weakening patent defenses	As core patent amendments narrowed the scope of rights, room for bypass designs has opened. The outcome of additional litigation is key.	Check for acceleration of competitor entry. Conservative revaluation.
③ HBF mass production delay	HBF is an early-stage technology aiming for 2027 mass production. If the schedule slips, pre-reflected expectations will evaporate.	Adjust the 'Super Bull' timeline backward. Base and Bull cases remain valid for 2nm and HBM.
④ Delay in new fab orders	If the CAPEX cycles of clients (Samsung, SK, Micron, TSMC) are delayed, equipment delivery and revenue recognition will also be delayed.	Monitor quarterly orders and revenue recognition.
⑤ Reversal of valuation overheating	A 2026E PER of 59x and PBR of 23x are vulnerable to earnings misses. Applying EPS × 40x results in the 50,000 KRW range.	Manage volatility by taking partial profits at 92,000 and 110,000 KRW.

Closing. HPSP is one of the world's quietest monopolies — the name may be unfamiliar, but every factory making state-of-the-art chips cannot move an inch without this company's equipment. 2nm, HBM, and new fabs have opened the harvest season for that monopoly, and HBF adds a new field called 'NAND front-end process' on top of it. Thus, the path is open beyond 110,000 KRW up to 150,000–200,000 KRW — provided that the path is followed only by passing the two checkpoints of 'HBF confirmed by numbers' and 'maintenance of the monopoly.' Since today's stock price has already bought a significant portion of that future, one should not act out of greed, but walk with discipline — confirming triggers, selling into strength, and buying into weakness.

This report is a narrative analysis summarizing the research and multiple discussions of the Goguma Team (Yangpa, Dotori, Ttangkong, Danggeun, Goguma).
The suggested target prices are scenarios based on securities firm consensus and the latest industry trends, and do not constitute investment advice. Investment decisions and responsibilities rest solely with the investor.
As of June 15, 2026 · HPSP (403870 · KOSDAQ)