PSK Inc. makes the machine that strips photoresist off a silicon wafer after it has been patterned β an unglamorous step that happens dozens of times per chip. Its own Korean filing prints a world market share of 40%. The same filing prints something the share number hides: nearly half of what PSK sells is not a machine at all, and three-quarters of last year’s growth came from that half.
π Key Takeaways
- PSK Inc. (KOSDAQ: 319660) is a 343-person Korean toolmaker whose filing claims 40% of the world dry strip market. Two filings print two different share series. The FY2025 annual report gives 26 / 42 / 38 / 40 / 40 for 2020–2024 (Gartner, April 2025); the half-year report filed five months later gives 42 / 38 / 40 / 40 / 39 for 2021–2025 (Gartner, March 2026). On the newer table the share peaked in 2021 and slipped in 2025 (Source: DART annual report, rcpNo 20260323000017, filed 23 Mar 2026).
- Only 51.8% of first-half 2026 revenue was equipment. The other 48.2% β $116.1M β is the line the filing labels “parts and service fees and other.” That share has climbed from 37.3% in FY2023 to 48.2% in H1 2026.
- Half-year operating profit more than doubled: $32.1M to $72.4M, a margin of 30.1%. PSK’s own Value-Up plan commits only to “maintain an operating margin above 15%” through 2030. It is running at twice its own target (Source: DART half-year report, rcpNo 20260814000731, filed 14 Aug 2026).
- China revenue halved in FY2025 β and the company still grew 14.9%. China fell from $127.5M to $68.4M while Korea rose 79%. One customer worth 21.3% of FY2024 revenue dropped below the 10% disclosure threshold entirely.
- Foreigners’ direction depends entirely on the window. Net buyers of $60.7M over 20 sessions to 4 Sep 2026 β net sellers of $116.9M over the 55 sessions we hold. Institutions did the exact opposite in both windows.
- PSK grades itself in public, and gave itself one failing mark. Its April 2026 Value-Up implementation filing scores four targets: revenue, margin and dividend all beaten; ESG rating target of “B or better” missed with a KCGS grade of C.
Correction, 8 September 2026. Within hours of publishing we found that PSK’s half-year report β a document this article already cites for revenue β prints a newer market-share series than the annual report we quoted, running one year further to 39% in 2025. The original text described the share as flat; on the newer table it slipped. The passages below now carry both series and name the filing each comes from.
π’ What Is PSK Inc., and What Does It Actually Sell?
PSK Inc. is a 343-person Korean company in Hwaseong that builds front-end semiconductor process tools β chiefly dry strip machines, which use plasma to remove the photoresist mask after a wafer has been patterned β and sells them directly to chipmakers in Korea, China, Taiwan, the US and Japan.
The process itself is old and universal. Every layer of a modern chip is patterned by coating the wafer in a light-sensitive polymer called photoresist, exposing it, etching through the gaps, and then removing whatever resist is left. That removal step is plasma ashing β in Korean filings and in PSK’s own English marketing, “dry strip.” It happens dozens of times in the making of a single logic or memory die, which is why a dry strip tool is a volume purchase rather than a marquee one.
The filing’s own product list is broader than that one word suggests. Under “principal business,” the FY2025 annual report names four families: Dry Strip, Selective Material Removal, New Hard Mask Strip and Bevel Etch, and states the company is “expanding its scope in order to become a global leader in comprehensive process equipment” (Source: DART annual report, rcpNo 20260323000017). Turn to a later section of the same annual report β the segment table headed μ¬μ λΆλ¬Έλ³ μ£Όμ μ ν β and the list changes: Dry Strip, Dry Cleaning, New Hard Mask Strip, Wafer Edge Clean and Power Device Etch. Five names, three of which did not appear a few sections earlier, and two of which (Selective Material Removal, Bevel Etch) have dropped out. The half-year report’s equivalent table gives a third set, shorter again: Dry Strip, Dry Cleaning, Wafer Edge Clean. So PSK describes its own product range three different ways in two filings, and no two lists match. That is a reminder that these labels are marketing categories rather than accounting ones β because underneath all three, PSK reports as a single segment. There is no product-level profit disclosure anywhere in the filings, and no reconciliation of the three lists either.
| PSK Inc. β the basics | As filed |
|---|---|
| Legal name | νΌμμ€μΌμ΄ μ£Όμνμ¬ / PSK INC. |
| Ticker | KOSDAQ 319660 |
| Incorporated | 1 April 2019, by spin-off (μΈμ λΆν ) from PSK Holdings |
| Listed | 10 May 2019 (KOSDAQ re-listing) |
| Head office & plant | 48 Samsung 1-ro 4-gil, Hwaseong, Gyeonggi β plus a Pangyo campus opened May 2024 |
| Employees | 343 (267 men, 76 women), average tenure ~4 yrs 10 mths, as of 31 Dec 2025 |
| Overseas footprint | 5 foreign subsidiaries (US, Taiwan, China Γ2, Japan) + 9 foreign branches |
| Patents | 291 registered, 311 filed (602 total), 31 Dec 2025 |
| Co-CEOs | Lee Kyung-il and Ha Hyung-chan (κ°μλν, each with independent authority) |
| Board chair | Park Kyung-soo (b. Dec 1952), who is also CEO of PSK Holdings |
Source: DART annual report rcpNo 20260323000017 (FY2025), sections I and VIII. Employee count is the separate-entity (λ³λ) figure for PSK Inc. itself, not the consolidated group. The Co-CEO row is the one exception, and it comes from the half-year report (rcpNo 20260814000731, as of 30 Jun 2026), which is where the κ°μλν designation appears: the annual report, dated 31 December 2025, records Lee Kyung-il as the sole representative director and notes Ha Hyung-chan’s appointment as an item still to be put to the 30 March 2026 AGM.
One fact worth stating plainly, because it shapes everything a foreign investor can find: PSK Inc. has no English-language Wikipedia entry. We checked PSK_Inc., PSK_Inc and PSK_Holdings on 7 September 2026; all three return 404. A company that its own filing credits with 40% of a world market has no encyclopedia page in the language most of its potential shareholders read. That is not a claim about the company. It is a description of the information gap this article exists to cross.
π Where Does PSK’s Revenue Actually Come From?
Just over half of it β 51.8% in H1 2026 β comes from selling machines; the rest comes from a line the filing calls “parts, service fees and other,” and that line has grown from 37.3% of revenue in FY2023 to 48.2% today.
| Revenue by type (consolidated) | FY2023 | FY2024 | FY2025 | H1 2026 (λμ ) |
|---|---|---|---|---|
| Equipment (μ ν Β· λ°λ체 곡μ μ₯λΉ λ₯ μΈ) | $163.8M | $171.0M | $182.1M | $124.6M |
| Parts, service fees & other (κΈ°ν Β· λΆν λ° μ©μμμλ£ μΈ) | $97.6M | $124.7M | $157.6M | $116.1M |
| Total | $261.4M | $295.7M | $339.7M | $240.8M |
| Non-equipment share of revenue | 37.3% | 42.2% | 46.4% | 48.2% |
Korean row labels reproduced from the filings. FY figures are full-year consolidated; the H1 2026 column is the cumulative (λμ ) six months to 30 June 2026, not the three-month quarter. Sources: DART annual report rcpNo 20260323000017, section II-4-κ° (λ§€μΆμ€μ ); DART half-year report rcpNo 20260814000731, same section. FX: β©1,346/$1 as of 4 Sep 2026 β applied to every dollar figure in this article. Percentages are computed on the underlying won; dollar columns are converted afterwards, so a column may miss its printed total by $0.1M through rounding.
Two things about that “other” line deserve care. First, the filing does not break it down further, so we cannot tell you how much is spare parts, how much is field service, and how much is the “other” in the label. Second, it is overwhelmingly an export line: in H1 2026, $107.8M of the $116.1M was booked as export and only $8.3M as domestic. In FY2025 the same split was $128.1M export against $29.4M domestic. PSK’s installed base outside Korea is where this money comes from.
Geographically, FY2025 was not a normal year. The segment note discloses revenue across five buckets, and two of them changed places:
| Revenue by region (consolidated) | FY2024 | FY2025 | Change | H1 2025 | H1 2026 |
|---|---|---|---|---|---|
| Korea (λνλ―Όκ΅) | $58.6M | $105.1M | +79.4% | $54.6M | $68.9M |
| China (μ€κ΅) | $127.5M | $68.4M | β46.3% | $39.5M | $51.0M |
| United States (λ―Έκ΅) | $27.2M | $26.5M | β2.6% | $13.5M | $34.6M |
| Taiwan (λλ§) | $31.2M | $43.4M | +39.2% | $16.6M | $32.3M |
| Other countries (κΈ°ν κ΅κ°) | $51.3M | $96.3M | +87.7% | $32.2M | $54.0M |
| Total | $295.7M | $339.7M | +14.9% | $156.3M | $240.8M |
The filing groups the last four columns under a single “μΈκ΅” (foreign) header; we have kept the four country buckets it actually prints. Sources: rcpNo 20260323000017, consolidated note 32 (μμ λΆλ¬Έ μ 보), and rcpNo 20260814000731, consolidated note 31. FY columns are full years; H1 columns are cumulative six-month figures. Region columns are converted from won individually and may miss the printed total by $0.1M through rounding.
Note the reversal inside the reversal. China halved across FY2025 β and then grew 29.2% year on year in the first half of 2026. Anyone who read only the annual report would conclude PSK had lost China; anyone who read only the half-year report would never know it had fallen. The two filings tell different stories about the same country, and both are true of their own window.
ποΈ How Did PSK Get Here?
PSK Inc. as a legal entity is only seven years old β it was carved out of PSK Holdings in April 2019 β but it inherited a dry strip business and a development pipeline that the filings trace back to at least 2012.
This matters for anyone looking at a long price chart. The KOSDAQ listing that carries PSK’s pre-2019 history belongs to the other ticker (see the ticker section below). PSK Inc.’s own record starts in May 2019.
| Date | Event, as recorded in the filing |
|---|---|
| Oct 2012 | 450mm dry strip development programme begins (ran to Dec 2025), inherited from the pre-split company |
| Apr 2019 | PSK Inc. incorporated by spin-off; the parent simultaneously renames itself PSK Holdings |
| May 2019 | KOSDAQ re-listing |
| 2020 | Designated a materials-parts-equipment specialist company; core strategic technology designation; Gold Tower Industrial Medal on Semiconductor Day |
| Jan 2022 | Designated a “μλΆμ₯ μΌλΈκΈ°μ ” (leading materials/parts/equipment firm) by the Korean government |
| Jul 2022 | Board cancels 281,663 treasury shares (μ΄μ΅μκ°) |
| Sep 2022 | 1-for-1 bonus issue of 14,483,357 shares, capitalising the share-premium account |
| May 2024 | Pangyo campus opens |
| Apr 2025 | First Value-Up (κΈ°μ κ°μΉμ κ³ κ³ν) plan filed |
| MarβApr 2026 | 2026 Value-Up plan filed (rcpNo 20260330901878), followed by a self-scored implementation review (rcpNo 20260430900483) |
| 7β8 Sep 2026 | Investor meetings in Hong Kong at Macquarie’s NEXT INNOVATIONS Tech Conference, targeted at overseas institutions (rcpNo 20260904900080, filed 4 Sep 2026) |
Compiled from the μ°ν (corporate history), μκΈ°μ£Όμ and μ¦μ tables of rcpNo 20260323000017 plus the individual disclosures linked above. The IR notice contains no financial figures β it is a schedule, not guidance.
π Does PSK Really Have 40% of the World’s Dry Strip Market?
Its FY2025 annual report prints exactly that β a 40% world market share for 2024, sourced to Gartner β alongside a five-year series showing the share peaked at 42% in 2021. But the half-year report filed on 14 August 2026 prints a different five-year window, one year further forward and re-sourced to Gartner’s March 2026 data: 42% (2021), 38%, 40%, 40%, and 39% for 2025. Neither filing states what the denominator is. The two tables overlap on four years and agree on all four; the newer one simply adds the year the older one could not have.
Korean annual reports rarely print market share by year. This one does, and the whole series is worth reproducing rather than the single most flattering point:
| PSK’s world market share, as printed | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|
| μΈκ³μμ₯ μ μ μ¨ (%) | 26 | 42 | 38 | 40 | 40 |
Reproduced in full, all five columns the filing prints. Source cited in the filing: “Gartner, 2025.04.” From rcpNo 20260323000017, section II-7-κ°-(4). The series stops at 2024; the FY2025 report does not print a 2025 share.
The sentence immediately above that table reads: “The Company is consolidating its world market share in the dry strip equipment field, and continues to work to expand it further.” Set that against the table. Share rose once, in 2021, and has moved by two points in the four years since. The prose says expanding; the table says flat. That is management’s characterisation, not a finding β and the table beneath it is the check.
The filing also prints the market PSK is taking share of, from a different Gartner cut:
| World dry strip market ($M) | 2022 | 2023 | 2024 | 2025F | 2026F | 2027F |
|---|---|---|---|---|---|---|
| Market size | 799 | 628 | 709 | 754 | 793 | 819 |
Source cited in the filing: “Gartner, 2025.02.” Same section of rcpNo 20260323000017. F = forecast. Dollar figures are the filing’s own; we have not converted them.
Quick Take β the denominator the filing never gives you. Put the two tables together and the arithmetic does not settle. PSK’s entire FY2024 revenue, converted at this article’s single rate, was $295.7M β which is 41.7% of Gartner’s $709M dry strip market for that year. FY2023: $261.4M against a $628M market is 41.6%. Both land within two points of the printed 40%. That is suspiciously tidy for a company where, on its own numbers, equipment was only $171.0M of FY2024 revenue β 24% of that market, not 40%. Either Gartner’s market definition is wider than “tools,” or the share is measured on something other than revenue, or the near-match is coincidence. The filing footnotes neither denominator, so we cannot tell you which. The ratio is also sensitive to the exchange rate used, and we are applying a September 2026 rate to 2023 and 2024 revenue. Treat the 40% as a company-supplied, Gartner-sourced figure we could not reconcile β not as a verified fact.
Two further disclosure gaps belong here, because both are the filing declining to answer rather than us failing to look. PSK does not disclose capacity utilisation: “because equipment is produced to order, capacity can be calculated but its correlation with actual production is low, so disclosure is omitted.” And PSK does not disclose order backlog: “in compliance with customer NDAs and company sales policy, detailed order status is omitted.” A reader who wants forward visibility on this company will not get it from DART.
π Is the Business Getting Better or Worse?
Sharply better in 2026, after a FY2025 that looked strong on revenue and flat at the bottom line β half-year revenue rose 54.0% and half-year operating profit rose 125.6%, lifting the operating margin to 30.1% from 19.4% for FY2025 as a whole.
| Consolidated (μ°κ²°) | FY2023 | FY2024 | FY2025 | H1 2026 (λμ ) | H1 2025 (λμ ) |
|---|---|---|---|---|---|
| Revenue | $261.4M | $295.7M | $339.7M | $240.8M | $156.3M |
| Operating profit | $40.2M | $62.4M | $65.7M | $72.4M | $32.1M |
| Operating margin | 15.4% | 21.1% | 19.4% | 30.1% | 20.5% |
| Pre-tax profit | $49.0M | $72.4M | $68.7M | $78.8M | $32.0M |
| Net profit attributable to owners | $39.0M | $58.8M | $58.4M | $63.1M | $28.7M |
| Effective tax rate | 20.3% | 18.8% | 15.1% | 19.9% | 10.4% |
All figures consolidated (μ°κ²°). FY columns are full years from rcpNo 20260323000017, consolidated statement of comprehensive income. H1 columns are the cumulative (λμ ) six-month columns of rcpNo 20260814000731 β not the three-month (3κ°μ) columns printed beside them in the same table. Effective tax rate = income-tax expense Γ· pre-tax profit, our calculation. FX β©1,346/$1 as of 4 Sep 2026.
The three-month column matters, because it is easy to mistake for the half. PSK’s half-year income statement prints four columns β Q2 and H1 for 2026, Q2 and H1 for 2025 β and the quarter alone is a large number. For the record: Q2 2026 (the 3κ°μ column) was revenue of $124.4M and operating profit of $37.3M, against $80.6M and $15.3M in Q2 2025. Both the quarter and the half doubled operating profit. Neither figure is the other one.
FY2025 was more complicated than its revenue line suggests. Revenue rose 14.9%, operating profit rose 5.4% β and net profit fell 0.7%, from $58.8M to $58.4M. The cause was below the operating line: non-operating income dropped from $10.0M to $3.0M as finance costs rose and finance income fell. A lower effective tax rate (15.1% against 18.8%) absorbed most of the rest. The clean way to describe FY2025 is that the operating business grew and the financial result shrank.
Rolling the two together gives a trailing-twelve-month picture that neither filing prints:
| Trailing twelve months to 30 Jun 2026 (our bridge) | Revenue | Operating profit | Net (owners) |
|---|---|---|---|
| FY2025 full year (as filed) | $339.7M | $65.7M | $58.4M |
| less H1 2025 (as filed) | β$156.3M | β$32.1M | β$28.7M |
| = H2 2025 (derived) | $183.4M | $33.7M | $29.7M |
| plus H1 2026 (as filed) | +$240.8M | +$72.4M | +$63.1M |
| = TTM to 30 Jun 2026 | $424.1M | $106.0M | $92.8M |
Our arithmetic, from the two filings named above. The H2 2025 row is derived by subtraction, not disclosed β PSK does not publish a second-half statement. TTM operating margin = 25.0%. Rounding in the dollar column can leave one-decimal differences against the won originals.
π Why Did Three-Quarters of Last Year’s Growth Come From the Line That Isn’t Equipment?
Because PSK shipped fewer machines at lower average prices in FY2025 than in FY2024, and revenue still grew 14.9% β 74.8% of that increase came from parts and services rather than from tools.
Let us state that precisely, because the arithmetic is the point. FY2025 revenue rose by $44.0M. Split by line:
| Where FY2025’s growth came from | FY2024 | FY2025 | Increase | Share of growth |
|---|---|---|---|---|
| Equipment | $171.0M | $182.1M | +$11.1M | 25.2% |
| Parts, service fees & other | $124.7M | $157.6M | +$32.9M | 74.8% |
| Total | $295.7M | $339.7M | +$44.0M | 100% |
Our calculation from the λ§€μΆμ€μ table of rcpNo 20260323000017. Share-of-growth percentages are computed on the underlying won figures, then the dollar column is converted at β©1,346/$1 β so the dollar increments may not sum exactly to the total after rounding.
Now the two supporting facts, both printed in the same filing and both pointing the same way.
Unit output fell. PSK’s production table reports 160 tools in FY2023, 191 in FY2024, and 187 in FY2025 β a 2.1% decline in the year revenue grew 14.9%. (The first half of 2026 has already produced 149, which annualises well above any prior year.)
Average selling price fell too. The filing prints an average unit price for “semiconductor process equipment”: β©1,251M in FY2023, β©1,453M in FY2024, β©1,422M in FY2025, and β©1,382M in H1 2026 β two consecutive declines across three periods from the 2024 peak, a cumulative 4.9%. The filing footnotes that specifications vary by customer order, so this is a mix-weighted average rather than a price list.
Fewer tools, cheaper tools, and revenue up 15%. The gap is the aftermarket line. That is a materially different business model from the one the phrase “semiconductor equipment maker” conjures β it is closer to an installed-base annuity with a capital-equipment front end. We want to be careful about how far to push this: PSK reports as a single segment and gives no profit split by line, so we cannot tell you whether parts and service carry a higher margin than tools. The margin did expand β 19.4% for FY2025 to 30.1% in H1 2026, in a period when the non-equipment share also rose β but that is a coincidence in the data, not a demonstrated cause. Cost, scale, mix and currency all moved at once, and the filings do not decompose them.
Underneath that, FY2025 also saw the customer base change shape. The segment note lists customers accounting for 10%+ of consolidated revenue, anonymised as A, B, C, D and E:
| Customers β₯10% of consolidated revenue | FY2024 | FY2025 | H1 2025 | H1 2026 |
|---|---|---|---|---|
| Customer A | $69.3M (23.4%) | $110.7M (32.6%) | $59.8M (38.3%) | $68.3M (28.4%) |
| Customer B | $32.9M (11.1%) | $57.6M (17.0%) | $20.8M (13.3%) | $46.5M (19.3%) |
| Customer C | $41.8M (14.1%) | $42.8M (12.6%) | $17.7M (11.3%) | $38.3M (15.9%) |
| Customer D | β (below threshold) | $42.6M (12.5%) | $21.5M (13.8%) | $32.9M (13.7%) |
| Customer E | $63.0M (21.3%) | β (below threshold) | β | β |
| Disclosed customers, total | 70.0% | 74.7% | 76.7% | 77.3% |
Sources: rcpNo 20260323000017 note 32 and rcpNo 20260814000731 note 31. β οΈ The letters are not guaranteed to refer to the same company across filings or across periods β DART does not say. The half-year note lists only four letters, A to D. Percentages are our calculation against total consolidated revenue for the matching period.
The filing’s own footnote explains the two changes: “Customer D was below the 10% threshold in the prior period and therefore not a disclosure subject; Customer E was a disclosure subject in the prior period but was not disclosed in the current period as its revenue share fell below 10%.”
So a customer worth 21.3% of FY2024 revenue β $63.0M β fell below $34.0M in FY2025 and out of the table. In the same year, China revenue fell by $59.1M. We are not asserting those are the same thing. The filing anonymises customers and does not link them to regions, and we have no independent source that does. What we can say is that both movements are in the same filing, are of similar size, and run in the same direction β and that PSK’s disclosed customer concentration nonetheless rose, from 70.0% to 74.7% and then to 77.3% in the latest half. Losing the largest departing customer did not diversify this company.
Elsewhere on our site we have written about a different flavour of the same problem β Dongjin Semichem selling its China business with Samsung as most of what remains, and HPSP, whose own filing says its market share cannot be calculated. Concentration and unverifiable share claims are close to universal in this corner of the Korean supply chain.
π― What Did PSK Promise Its Own Shareholders β and Did It Deliver?
PSK filed a Value-Up plan with four numeric targets for 2026β2030 and then filed a public self-assessment against them: it beat three and failed one, the ESG rating.
Korea’s Value-Up programme asks listed companies to publish voluntary corporate-value plans. Most of what gets filed is aspirational prose. PSK’s is unusually specific, and β rarer still β it comes back a month later and marks its own homework.
| PSK’s own Value-Up target (2026β2030) | PSK’s own score against FY2025 | Verdict |
|---|---|---|
| Revenue CAGR of 10% | “Revenue grew 14.9%” | Met |
| Operating margin maintained above 15% | “Operating margin 19.4%” | Met |
| Dividend per share maintained at β©500 or more | “Dividend β©680” | Met |
| KCGS overall ESG rating of B or better | “KCGS overall rating C“ | Missed |
Targets from the Value-Up plan filed 30 Mar 2026 (rcpNo 20260330901878); scores from the implementation review filed 30 Apr 2026 (rcpNo 20260430900483), which evaluates the plan first filed 30 Apr 2025. Quotations are our translation of the Korean text; the “Verdict” column is ours, reading the company’s own scores against its own targets. KCGS = Korea Institute of Corporate Governance and Sustainability. Both filings carry Korea’s standard forward-looking-information caveat.
Two observations. First, the margin target is remarkable in the other direction: PSK committed to “above 15%” and delivered 19.4% for FY2025 and 30.1% in H1 2026 β double its own floor. A target a company clears by 100% is not a stretch goal; it is a floor set low enough to be safe. Second, and to the company’s credit, it published the failing grade. A Value-Up filing that only reported the three wins would have been entirely within the rules. This one prints the C.
The same filing also records PSK as qualifying as a “high-dividend company” under Article 104-27 of Korea’s Restriction of Special Taxation Act β a status the filing’s own footnote says rests on “the company’s own judgement” (κΈ°μ μ μ체μ μΈ νλ¨μ μν¨), not on a regulator’s determination β with FY2025 dividends of $14.6M against FY2024’s $8.6M β a 70.0% increase, the figure the disclosure itself computes.
π° What Do You Actually Pay for PSK Today?
At the 4 September 2026 close of β©126,500 (~$94), PSK is a $2.72B company trading at 46.7Γ FY2025 earnings, 29.3Γ trailing-twelve-month earnings, and 6.0Γ book β after a 242.8% year-to-date move that still leaves it 40.2% below its 52-week high.
| PSK Inc. (319660) valuation snapshot | As of 4 Sep 2026 |
|---|---|
| Share price | β©126,500 (~$94) |
| Shares outstanding | 28,966,714 β zero treasury shares |
| Market capitalisation | $2.72B |
| P/E on FY2025 net profit | 46.7Γ |
| P/E on TTM net profit (to 30 Jun 2026) | 29.3Γ |
| Price/book (30 Jun 2026 equity) | 6.0Γ |
| Dividend per share (FY2025) / trailing yield | β©680 / 0.54% |
| Payout ratio (FY2025, consolidated) | 25.1% |
| Year to date | +242.8% |
| 52-week range / position | β©23,800 β β©211,500 Β· 40.2% below the high |
| Debt-to-equity (30 Jun 2026) | 22.06% (the filing’s own ratio) |
Price, share count and market cap from the Korea Exchange settled close of 4 Sep 2026 β a settled close, not an intraday quote. Earnings and equity from the DART filings cited above; net profit is μ§λ°°μ£Όμ£ΌλΉκΈ°μμ΄μ΅ (attributable to owners of the parent). TTM P/E uses the bridge computed earlier in this article. Book value uses 30 Jun 2026 consolidated equity less the β©108,827,548 non-controlling interest. FX β©1,346/$1 as of 4 Sep 2026. The 52-week range is measured on closing prices over the 243 sessions from 5 Sep 2025 to 4 Sep 2026 β and we checked the low end against the company’s own filing rather than a data feed, because a 52-week window is easy to cut by row count instead of by date. The annual report’s monthly price table prints a September 2025 low of β©23,050 and an August 2025 low of β©19,410, so any low near β©20,000 belongs to August β outside a window that opens on 5 September. Every figure in this table goes stale β check a live quote before acting on any of it.
Quick Take β the dividend yield has been priced away. The FY2025 annual report prints a cash dividend yield of 1.80%, against 2.40% for FY2024 and 1.00% for FY2023. That is not the yield a buyer gets today. The disclosed figure is computed on the filing’s own price basis for the dividend year; on the 4 September 2026 close of β©126,500, the same β©680 dividend is 0.54%. PSK tripled its dividend per share in two years β β©200 to β©400 to β©680 β and the yield still fell, because the share price rose faster than the payout. Always check which price a Korean filing’s λ°°λΉμμ΅λ₯ is divided by before comparing it to a live quote.
Against Korean peers, PSK sits in the middle of the sector on price and near the bottom on margin:
| Korean semiconductor equipment peers | Market cap | FY2025 revenue | Operating margin | P/E | Basis |
|---|---|---|---|---|---|
| HPSP (403870) | $3.11B | $129M | 52.0% | 57.5Γ | λ³λ OFS |
| Hanmi Semiconductor (042700) | $16.29B | $428M | 43.6% | 102.4Γ | μ°κ²° CFS |
| EO Technics (039030) | $4.05B | $283M | 21.2% | 94.8Γ | μ°κ²° CFS |
| PSK Inc. (319660) | $2.72B | $340M | 19.4% | 46.7Γ | μ°κ²° CFS |
| Eugene Technology (084370) | $2.32B | $260M | 14.7% | 70.3Γ | μ°κ²° CFS |
| Jusung Engineering (036930) | $6.16B | $231M | 10.1% | 232.2Γ | μ°κ²° CFS |
β οΈ Not like-for-like: HPSP reports separate (λ³λ) accounts because it has no subsidiaries; every other row is consolidated (μ°κ²°). We have labelled the basis rather than blending it. P/E = market cap at the 4 Sep 2026 close Γ· FY2025 net profit; PSK’s net profit is μ§λ°°μ£Όμ£ΌλΉκΈ°μμ΄μ΅, the others are λΉκΈ°μμ΄μ΅. This table is reproduced unchanged from our HPSP article of 6 September 2026, so the two pieces cannot disagree with each other. Sources: DART FY2025 annual reports via the structured accounts API; KRX closes. FY2025 figures are stale relative to PSK’s H1 2026 β on TTM earnings PSK’s multiple is 29.3Γ, and we have not recomputed the peers on the same basis.
π·οΈ Which PSK Ticker Are You Actually Buying?
There are two listed companies called PSK on KOSDAQ, and they make different things: 319660 is the front-end tool business, while 031980 (PSK Holdings) makes back-end packaging equipment and separately owns 32.76% of 319660 directly.
This is the single most likely way for a foreign investor to buy the wrong security here. The confusion has a specific origin: the original listed company was called PSK Inc. On 1 April 2019 it renamed itself PSK Holdings and spun out a new PSK Inc. So a long price chart labelled “PSK” β and any pre-2019 financial history β belongs to 031980, not to the ticker most people mean today.
| Entity | Ticker | What it makes | Market cap / FY2025 revenue |
|---|---|---|---|
| PSK Inc. (νΌμμ€μΌμ΄) | KOSDAQ 319660 | Front-end (μ 곡μ ): dry strip, selective material removal, new hard mask strip, bevel etch | $2.72B / $339.7M |
| PSK Holdings (νΌμμ€μΌμ΄νλ©μ€) | KOSDAQ 031980 | Back-end packaging (ν곡μ ): descum, fluxless bump reflow, hot DI water β plus its 32.76% direct stake in 319660 | $2.31B / $154.4M |
| PSK Holdings (Pyeongtaek entity) | Unlisted, absorbed | The former unlisted holding vehicle, merged into the listed 031980; this is why the filing records a change of largest shareholder from one “PSK Holdings” to another | β |
| PSK International, and a second unlisted “PSK” | Unlisted | Separate DART registrants sharing the name; not the listed company | β |
| SEMIgear, PSK America, PSK Asia, PSK Japan, PSK China / Shanghai | Unlisted | Group affiliates. SEMIgear (Texas) sits under PSK Holdings; the five regional sales/service arms sit under PSK Inc. β PSK America (Austin), PSK Asia (Taiwan), the Japanese entity (Hiroshima), PSK China (Xi’an) and PSK Shanghai | β |
Entity list from DART’s corporate-code register (checked 7 Sep 2026) and the affiliate table of rcpNo 20260323000017 section I-1-λ§ (13 group companies, 11 unlisted). PSK Holdings’ product description is from its own half-year report, rcpNo 20260814000743. Market caps at the 4 Sep 2026 KRX close.
There is a valuation consequence to that structure worth doing the arithmetic on. PSK Holdings’ own 9,489,994 shares of PSK Inc. β the reporting entity’s line in the large-holding filing, before the group executives who are counted alongside it β were worth $891.9M at the 4 September close. PSK Holdings’ own market capitalisation was $2.31B. So the stake in the listed subsidiary accounts for 38.7% of the parent’s market value; the remaining $1.41B is the market’s price for a packaging-tool business that did $154.4M of FY2025 revenue at a 35.3% operating margin. We have not opened PSK Holdings’ full accounts for this article and are not offering a view on whether that is cheap or dear β but a reader comparing the two tickers should know that buying 031980 is roughly 61% a bet on back-end packaging tools and 39% a look-through claim on 319660.
We found no American Depositary Receipt for either ticker. Neither appears in our list of Korean stocks available as US-listed ADRs. Foreign buyers use the Korean line β see how to buy Korean stocks as a foreign investor, and how won prices convert to dollars. That is a statement about where we looked, not a guarantee that no instrument exists anywhere.
π Is Foreign Money Buying or Selling PSK?
Both β and which answer you get depends entirely on the window you pick: over the twenty sessions to 4 September 2026 foreigners were net buyers of PSK Inc. (319660), over the fifty-five sessions we hold they were net sellers, and Korean institutions took the opposite side in each. The table below gives both windows rather than the flattering one.
| PSK Inc. 319660 β net flow to 4 Sep 2026 | 5 sessions | 20 sessions | 55 sessions |
|---|---|---|---|
| Window covered | 31 Aug β 4 Sep | 7 Aug β 4 Sep | 18 Jun β 4 Sep |
| Foreign investors (319660) | +252,486 sh +$23.3M |
+654,181 sh +$60.7M |
β403,581 sh β$116.9M |
| Korean institutions (319660) | β210,955 sh β$19.6M |
β614,156 sh β$57.4M |
+498,157 sh +$123.6M |
| Retail (319660) | β29,574 sh | β31,970 sh | β117,605 sh |
Flows arrive as share counts; the dollar figures are derived β each day’s net shares multiplied by that day’s close, summed, then converted at β©1,346/$1. That is an approximation, not execution prices. The 55-session column is the full length of our own settled-flow history for this ticker (from 18 Jun 2026), not a standard window β we label it that way rather than call it “three months.” Source: KIS Open API into our own investor_flows table, settled closes only. These are net trading flows, not a holdings level β we are not quoting an aggregate foreign ownership percentage here, because the sources that publish one for a KOSDAQ stock are not ones a reader can independently check.
The sign flips, so the headline flips. A story built only on the twenty-session column reads “foreigners are accumulating PSK.” A story built only on the fifty-five-session column reads “foreigners are exiting PSK.” Both would be arithmetically correct and both would be misleading on their own. What the two columns together actually say is narrower and more useful: foreign investors sold into the earlier part of the summer and have been buying back over the last month, while domestic institutions did the reverse in each window and remain net buyers across the full period. Nobody in this table is a stable holder.
The price does not obviously reward either side. Across the 55-session window in which foreigners were net sellers, PSK fell 26.5% β the institutions were buying into that decline. Across the 20 sessions in which foreigners were net buyers, the stock rose 6.8%, and the institutions were selling into it. We are describing sequence, not causation: with two large counterparties on opposite sides of every one of these windows, no flow column here explains the price on its own.
The institutional side shows up in the filing record too. Three 5%-plus positions were reported in 2026 filings β one of them the controlling shareholder β and the FY2025 annual report footnotes two more on older basis dates:
| Holder | Shares | Stake | Filing |
|---|---|---|---|
| PSK Holdings + related parties | 9,513,482 | 32.84% | 6 Jul 2026 β 9,514,112 β 9,513,482 shares after a related party sold 630; the filing rounds both to 32.84% |
| Shinhan Asset Management | 1,835,045 | 6.34% | 20 Jul 2026 β new 5% filing, purpose “simple investment” |
| Samsung Asset Management | 1,721,504 | 5.94% | 8 Jul 2026 β down from 7.30% |
| National Pension Service | 2,357,903 | 8.14% | Basis date 29 Dec 2025, per rcpNo 20260323000017 |
| Barings Asset Management | 1,753,606 | 6.05% | Basis date 1 Oct 2025, per rcpNo 20260323000017 |
β οΈ Different basis dates β do not add these up. The top three rows are 2026 filings; the bottom two are the basis dates the FY2025 annual report itself footnotes, and both are older than the 2026 filings above them. Sources as linked, plus section VII (μ£Όμ£Όμ κ΄ν μ¬ν) of rcpNo 20260323000017. For context on how these flows are reported and why the windows matter, see our Foreign Flow Watch series.
ποΈ Who Controls PSK, and What Do Minority Holders Get?
PSK Holdings, itself listed on KOSDAQ, controls 32.84% including related parties β and minority holders get a rising cash dividend and nothing else: PSK holds zero treasury stock and has cancelled no shares since 2022.
Be careful with the control number, because the filings give two. The FY2025 annual report’s shareholder table shows PSK Holdings alone at 9,489,994 shares, or 32.76%, as of 31 December 2025. The July 2026 large-holding report β the one that must include related parties β shows 9,513,482 shares, or 32.84%. The 23,488-share gap is nine individuals the filing lists as νΉλ³κ΄κ³μ, and it names their relationship to the reporting entity in a single repeated phrase: κ³μ΄νμ¬ μμ β officers of group companies. There are no family members on the list. Quote the wrong number and you either understate the control block or credit the parent with shares its executives hold personally; this is precisely the kind of two-number sentence Korean filings are full of, and the July filing prints both, one table apart.
The float is unusually wide for a Korean company at this level of control. The annual report records 28,439 small shareholders holding 17,687,209 shares β 61.07% of the register as of 31 December 2025. There is no cross-shareholding web here, no circular ownership, and no preferred share class: one common line, 28,966,714 shares, and zero treasury stock.
That last fact deserves emphasis in a Value-Up context. Korea’s reform push has centred on buybacks and, above all, on cancellation β retiring shares permanently to shrink the count. PSK did that once, in July 2022, retiring 281,663 shares; two months later it issued 14,483,357 new shares in a bonus issue that capitalised the share-premium account. Since then: no buyback, no cancellation, no treasury holdings at all. PSK’s entire shareholder-return policy is the dividend.
| Shareholder returns | FY2023 | FY2024 | FY2025 |
|---|---|---|---|
| Dividend per share | β©200 | β©400 | β©680 |
| Total dividend paid | $4.3M | $8.6M | $14.6M |
| Payout ratio (consolidated) | 11.0% | 14.6% | 25.1% |
| Cash dividend yield as disclosed | 1.00% | 2.40% | 1.80% |
| Buyback / cancellation | None | None | None |
Source: section III-6 (λ°°λΉμ κ΄ν μ¬ν) and section II of rcpNo 20260323000017, μ£Όμλ°°λΉμ§ν table, which prints νκΈλ°°λΉκΈμ΄μ‘ of β©5,793M, β©11,587M and β©19,697M for the three years. β οΈ The “as disclosed” yield row is the filing’s own νκΈλ°°λΉμμ΅λ₯ and is not computed on today’s price β see the Quick Take above. Payout ratio is on a consolidated basis, as the filing states.
On board structure, the filings record a chairman (Park Kyung-soo, born December 1952) who is simultaneously CEO of the controlling shareholder, PSK Holdings β and who holds no shares in PSK Inc. himself. The two co-CEOs, an inside director and the chairman hold cross-appointments across the group’s overseas subsidiaries. Outside directors were appointed in 2024 and 2025. The KCGS overall grade the company reports against itself is C.
β οΈ The Bear Case
- The growth target does not fit the market the filing describes. PSK commits to a 10% revenue CAGR through 2030. Its own Gartner table has the dry strip market going from $709M in 2024 to $819M in 2027 β about 4.9% a year β while its own share table has been flat at 40% for two years. Ten percent compound growth has to come from somewhere other than dry strip, and the filings do not say where.
- This article’s own central finding cuts both ways. We argue the aftermarket line is the story. But a business where units shipped fell (191 to 187) and average selling price fell (β©1,453M to β©1,382M over three periods) while parts revenue carried the growth is also a description of a company whose core tool franchise stopped growing. We cannot distinguish “durable annuity” from “tool sales stalled” with the disclosure available β and we have not tried to.
- R&D is shrinking as a share of revenue and in absolute terms. R&D was 8.7% of revenue in FY2024, 7.8% in FY2025 and 4.7% in H1 2026. Annualising the half’s $11.4M gives roughly $22.8M against FY2025’s $26.5M β a 14% cut in absolute spend, in a business the filing itself describes as having “very high technical barriers to entry.” Some of the current margin is R&D that was not spent.
- Customer concentration got worse, not better, after losing a 21% customer. The four disclosed customers were 70.0% of FY2024 revenue and 77.3% of H1 2026. The largest single customer alone was 28.4% of the latest half. PSK also cannot disclose backlog (customer NDAs) or utilisation (stated as uninformative), so a shareholder has no forward visibility on any of it.
- Governance: the controlling shareholder is a listed company whose own value depends on this one. PSK Holdings’ own 32.76% stake is 38.7% of its own market capitalisation. That is a structure with two sets of minority shareholders whose interests are not automatically aligned, and PSK Inc. holds zero treasury stock with which to defend its own share price. The company scores its ESG grade as C against its own B target.
- Earnings quality: FY2025’s bottom line already went backwards once. Net profit fell 0.7% in FY2025 despite 14.9% revenue growth, because non-operating income fell $7.0M and the tax rate did the rest of the work. In H1 2026 the effective tax rate went the other way β 10.4% to 19.9% β meaning pre-tax profit grew 146.5% while net profit grew 120.3%. Neither year’s headline profit is a clean read on operations.
- The price has already moved. The shares are up 242.8% year to date and the 52-week range of closing prices runs from β©23,800 to β©211,500 β a spread of roughly nine times inside twelve months. Whatever the business does next, the entry price is not the one that produced FY2025’s 46.7Γ multiple.
π Lingo Check
| Term | What it means | νκ΅μ΄ |
|---|---|---|
| dry strip / plasma ashing | Removing leftover photoresist from a patterned wafer using a plasma rather than a liquid chemical. Repeated dozens of times per chip, which makes the tools a volume purchase. | λλΌμ΄ μ€νΈλ¦½ |
| front-end vs back-end (μ 곡μ / ν곡μ ) | Front-end tools build circuits on the wafer; back-end tools cut, stack, bond and package the finished dies. PSK Inc. is front-end; PSK Holdings is back-end. The distinction is why two companies with the same name are not competitors. | μ 곡μ / ν곡μ |
| parts and service-fee revenue | The filing line κΈ°ν Β· λΆν λ° μ©μμμλ£ μΈ β spare parts, field service and other income from the installed base, reported separately from equipment sales but not broken down further. | λΆν λ° μ©μμμλ£ |
| three-month vs cumulative (3κ°μ / λμ ) | Korean half-year and quarterly statements print both a standalone-quarter column and a year-to-date column side by side. Reading the wrong one is the single most common error in translating a Korean filing. | 3κ°μ / λμ |
| spin-off (μΈμ λΆν ) | A horizontal split in which existing shareholders receive shares in the new entity pro rata, as opposed to λ¬Όμ λΆν , where the parent keeps 100% of the new company. PSK Inc. was created by μΈμ λΆν in April 2019. | μΈμ λΆν |
| Value-Up implementation review | The follow-up filing in which a company scores itself against the numeric targets in its earlier Value-Up plan. Voluntary, and rarely as specific as PSK’s. | κΈ°μ κ°μΉμ κ³ κ³ν μ΄ννν© |
| large-holding report (5% rule) | Korea’s disclosure requirement for anyone crossing 5% of a listed company, or moving 1% thereafter. The “μΌλ°” form is fuller than the “μ½μ” short form, and it includes related parties β which is why its stake number can differ from the annual report’s. | μ£Όμλ±μ λλ보μ μν©λ³΄κ³ μ |
| KCGS rating | The Korea Institute of Corporate Governance and Sustainability’s annual ESG grade, on a scale running S, A+, A, B+, B, C, D. PSK reports its own overall grade as C. | KCGS μ’ ν©λ±κΈ |
π― Why It Matters for K-Export Stars
Everything decisive in this article was in Korean and nowhere else. The five-year market-share series, the 51.8/48.2 revenue split, the falling unit count, the customer that dropped off the disclosure table, the Value-Up scorecard with its one failing grade β all of it sits in two DART filings that have no English version, filed by a company with no English Wikipedia page. The English-language record of PSK Inc. is a market-share number and a ticker.
That gap is the whole reason this site exists. Korea’s semiconductor supply chain is dense with companies like this one: small enough that no Western broker covers them, essential enough that the chipmakers everyone does cover cannot ship without them. Reading their filings is not exotic analysis. It is just reading β in the language they were written in.
Conclusion
PSK Inc. is a $2.72B Korean toolmaker whose own filing credits it with 40% of the world dry strip market, and whose own newer table shows that share slipping to 39% in 2025, against a market growing about 5% a year. Its recent growth has come from somewhere else: parts and service now make up 48.2% of revenue and supplied three-quarters of FY2025’s increase, while unit shipments and average selling prices both fell. The first half of 2026 was genuinely excellent β revenue up 54%, operating margin 30.1%, double the floor the company set for itself β and it arrived alongside a customer base that got more concentrated, an R&D budget that shrank, and a share price up 242.8% year to date.
None of that resolves into a recommendation, and we are not offering one. What it resolves into is a shorter list of the right questions: is the aftermarket line an annuity or a symptom; where does a 10% growth target come from in a 5% market; and does a company with zero treasury stock and a C-grade governance score have anything to give minority holders beyond a 0.54% yield. If you want to read further in this cluster, start with HPSP and Hanmi Semiconductor, then the Value-Up programme that PSK is grading itself against.
This company is one of ten in our K-Semiconductor Filing Map — a supply-chain map where every cell is marked by what the filing itself supports, and each one links to the receipt number it came from.
Disclaimer: This article is for informational and educational purposes only. It is not investment advice, an offer, or a solicitation to buy or sell any security. All figures are drawn from public regulatory filings and market data as of the dates stated and will go out of date; prices, multiples and flow figures in particular change daily. K-Export Stars holds no position in PSK Inc. or PSK Holdings. Do your own research and consult a licensed adviser before making any investment decision. See our corrections policy.
