Wonik IPS (240810): Korea’s Deposition Toolmaker, Decoded

Wonik IPS builds the machines that lay down the thin films inside a memory chip. In the first half of 2026 it booked $283.3m of revenue and $21.6m of operating profit — almost exactly what it earned a year earlier. Net profit, meanwhile, rose 388%. The gap between those two numbers is not a better business. It is a single block of 1,120,000 shares in another company, and by early September most of that gain had gone.

🏢 What Is Wonik IPS, and What Does It Actually Sell?

Wonik IPS is a Korean semiconductor and display capital-equipment maker that builds deposition and thermal-processing tools — the machines that grow the thin films inside a chip or a display panel — and it sells essentially all of them to a handful of Korean-headquartered customers.

The legal entity is young; the business is not. Wonik IPS Co., Ltd. (주식회사 원익아이피에스) was incorporated on 4 April 2016 as the spin-off entity carved out of Wonik Holdings’ semiconductor, display and solar-cell equipment division, with a demerger date of 1 April 2016. It relisted on the KOSDAQ market on 2 May 2016. (Source: DART annual report FY2025, filed 16 Mar 2026, rcpNo 20260316001453, Section I.) There is no English-language Wikipedia entry for the company that we could find, by direct lookup or by search; the Korean Wikipedia entry exists, which is itself a fair summary of how much of this company is documented only in Korean. The company’s own English IR site is the other primary reference.

The product line splits three ways in the filing, though the company reports as a single segment:

Everything is made to order. The filing is blunt that it cannot publish a price trend because “all sales are by fully customised build-to-order, and specification changes requested by the customer and changes to each tool’s chamber configuration mean the selling price can differ even for the same product.” That single sentence explains most of what follows: utilisation swings, lumpy revenue, and an order book that moves before revenue does.

Four sites carry the work: the Pyeongtaek head office and integrated R&D centre; a Jinwi industrial-complex plant for semiconductor tools; a Dunpo plant in Asan for display tools; and a Hwaseong R&D site. Headcount at 31 December 2025 was 1,509 — 1,467 permanent and 42 fixed-term — with average tenure of 7.6 years and average annual pay of ₩85,537,000 (~$63,500). Five consolidated subsidiaries, all unlisted, sit in China (Wuxi, Xi’an, Kunshan), the United States (Austin, Texas) and Singapore — every one of them registered as an equipment after-sales-service operation rather than a manufacturing site. Credit rating is A+ from eCredible, dated 5 September 2025. (Source: rcpNo 20260316001453.)

📊 Where Does Wonik IPS’s Revenue Actually Come From?

Roughly three-quarters is semiconductor equipment and roughly one-quarter is display, and about 71% of it is billed inside Korea — but the filing only splits the segments in the management-discussion section, not in the business section, which says the divisions cannot be separated at all.

This is worth pausing on, because it is a small example of why reading the Korean original matters. Section II.2 of the annual report states that “because products are manufactured within limited facilities using the same equipment and personnel, dividing the business into segments is not realistically possible, so this is prepared as a single business segment.” A reader who stops there concludes no segment data exists. Section IV — the directors’ management discussion — then prints the split by product and by region for two years.

By destination, the picture has been shifting for three years — and here the filing gives three points, not two:

The company’s own explanation for the FY2025 jump in overseas sales is specific and worth quoting rather than paraphrasing: overseas revenue rose 48.3% “on the back of increased equipment investment at the China plants of major domestic semiconductor customers, and increased OLED equipment investment at Chinese display panel customers.” In other words, a large part of the “export” line is still the same Korean customers — buying for fabs located in China. That distinction matters to anyone modelling this as a China-exposure story.

🗓️ How Did Wonik IPS Get Here? A Short Corporate History

Two corporate events define the company’s history: the 2016 spin-off that created it, and the 2019 absorption of Wonik Tera Semicon, which is why a ticker for that company no longer exists. The operating business is older than the legal entity, because it was carved out of a group founded decades earlier; the timeline below starts where the current company’s own filings start.

🌍 Is Wonik IPS a Global Leader, or Just a Korean Supplier?

The filing declines to answer: it states that estimating market share is impossible because total order volumes at chipmakers and per-supplier order volumes cannot be determined. The Korean sentence is “반도체 장비 제조 산업의 특성상 타 소자업체에 대한 총 발주량 및 업체별 수주량을 파악하기 어려운 관계로 실질적인 시장점유율 추정은 불가능 합니다” — “given the nature of the semiconductor equipment manufacturing industry, because it is difficult to ascertain total order volumes placed with other device makers and order volumes by company, a meaningful market-share estimate is impossible.” It says the same of the solar equipment market.

That is the same disclosure posture we found at HPSP, whose filing also says its market share cannot be calculated. It is common across Korean tool vendors, and it means any “Wonik IPS has X% of the ALD market” number you see is a third-party estimate, not a company-reported one.

What the filing does quantify is capacity and utilisation — and the three-year series is more interesting than the single year the company highlights:

Semiconductor nameplate capacity rose 41.4% over two years while units shipped went 198 → 178 → 195. Whatever else that says, it does not describe a plant running hot. The company’s own framing is that “the equipment business is an order-driven, build-to-order industry, so utilisation tends to vary widely with the order situation” — which is a fair caution against reading a low ratio as distress.

📈 Is the Business Getting Better or Worse?

On a full-year view, sharply better: revenue rose 21.6% and operating profit rose 593.6% in FY2025 after a loss-making FY2023. On the most recent six months, flat — and on the most recent three months, down.

Two things stand out. First, R&D is 100% expensed — the filing’s own table shows ₩0 capitalised as a development intangible in each of the three years. That is a conservative accounting choice and it means operating profit here is not flattered by capitalisation, unlike some peers. Second, R&D at $124.4m was 2.3× FY2025 operating profit. This is a business whose entire margin sits on top of a very large fixed research bill.

And then there is the order book, which is the one forward-looking number a build-to-order equipment maker actually discloses:

Backlog +112.8% in six months, on revenue up 4.10%, is the single most forward-looking fact in the report. It is also, unhelpfully, a separate-basis number sitting next to consolidated revenue, and there is no customer or delivery-date detail behind it. Treat it as a direction, not a forecast.

🔍 Why Did Net Profit Rise 388% While Operating Profit Rose 0.16%?

Because ₩89.3bn (~$66.4m) of the half-year’s pre-tax profit is an unrealised mark-to-market gain on financial instruments, and almost all of it comes from one 4.7% shareholding in another listed Korean chip-equipment maker.

Here is the H1 2026 income statement below the operating line, side by side with the year before. These are the filing’s own lines, cumulative for six months, consolidated:

What is VM Inc., and why does Wonik IPS own 4.7% of it?

VM Inc. (KOSDAQ 089970) is a Korean plasma-etch equipment maker, formerly named Adaptive Plasma Technology Corp., and Wonik IPS has held 1,120,000 of its shares since May 2023 as what the filing calls a “simple investment.”

The name change is why the position is easy to miss. Adaptive Plasma Technology (에이피티씨) became VM Inc. (브이엠) at its 28 March 2024 annual meeting, “to change the trade name for the company’s management objectives and business specialisation” (Source: DART, rcpNo 20240328902121). Wonik IPS’s own equity-investment schedule flags this in a one-line footnote: “the trade name was changed during the prior period.”

VM’s business is easy to place. In FY2025 it reported revenue of $107.3m and net profit of $19.4m (Source: DART consolidated financial statements, FY2025). In 2026 it disclosed two single-contract awards to SK hynix: ₩81,556,000,000 on 24 March 2026, which the filing itself scores at 116.05% of VM’s most recent annual revenue, and ₩21,578,000,000 on 23 June 2026 (Sources: rcpNo 20260324901398, rcpNo 20260623900188). Mind the base on that 116.05%: the filing’s “most recent revenue” line is FY2024 revenue of ₩70,277,411,560 ($52.2m), not the FY2025 figure quoted above — the disclosure says so explicitly, noting that “the most recent revenue of the company and of the counterparty is stated on the basis of the 2024 consolidated financial statements.” Against FY2025 revenue the same contract is 56.5%.

VM’s share price rose from ₩30,200 at the end of 2025 to ₩109,300 on 30 June 2026, up 262%. Wonik IPS owns 4.7% of the company — but note the date on that percentage. 4.7% is the figure Wonik IPS’s FY2025 annual report prints, as at 31 December 2025, and VM’s share count has grown since: 24,771,995 shares at 30 June 2026 against 26,326,765 on 4 September 2026. The same block of 1,120,000 shares is therefore roughly 4.3% of VM today. The share count is the constant here; the percentage is not. (Sources: stake percentage from the DART annual report cited above; prices and share counts are KRX settled data.)

The mark on that holding, as the filings print it:

The “implied per share” column is not in the filing; it is the carrying value divided by the share count. We computed it as a check, and it lands exactly on VM’s settled KRX close on each of those three dates. That confirms three things at once: the share count is right, the asset is marked to the quoted price with no discount, and the 30 June valuation is a clean market mark rather than a model.

The arithmetic of the half-year, then: the stake was carried at ₩33,824,000,000 on 31 December 2025 and ₩122,416,000,000 on 30 June 2026. That is ₩88,592,000,000 (~$65.8m) of increase, against a total reported “valuation gain on financial instruments” of ₩89,331,650,000 (~$66.4m). One holding is 99.2% of the line. Measured against the company’s own pre-tax profit for the same six months — the only base-consistent comparison, since the gain is a pre-tax figure — that single mark is 69.4% of ₩127,607,447,141, and 3.0× the ₩29,117,870,941 the equipment business earned at the operating line.

Does the underlying business look better if you strip the mark out?

Yes, but not because of operations — the improvement below the operating line is a currency swing. Take the valuation gain out of both years’ pre-tax profit and you get ₩38.28bn for H1 2026 against ₩23.08bn for H1 2025, an increase of ₩15.20bn. Over the same period the foreign-exchange lines in the “other income and expense” note swing by ₩16.62bn: net FX went from −₩8.88bn in H1 2025 (FX transaction losses ₩4.04bn plus translation losses ₩5.41bn against ₩0.57bn of gains) to +₩7.74bn in H1 2026 (₩3.01bn transaction gains plus ₩5.93bn translation gains against ₩1.21bn of losses). (Source: rcpNo 20260814001845, Note 28.)

In other words, the FX swing alone is larger than the entire ex-valuation-gain improvement in pre-tax profit. We are not asserting a mechanism for why the FX lines moved that way — the filing does not explain it, and the company’s foreign-currency asset and liability balances do not make the direction obvious. We are only noting that once you remove one non-cash equity mark and one currency swing, the six months look very much like the six months before: revenue up 4.10%, operating profit up 0.16%.

Readers who followed our work on SK hynix, where non-operating income matched operating profit in a record quarter, will recognise the shape. The difference of degree matters: at SK hynix the non-operating line roughly matched operations. Here it is more than three times operations, on a company one-fiftieth the size, and it is one position.

👥 Who Are Wonik IPS’s Customers, and How Concentrated Is It?

Three customers each accounted for more than 10% of consolidated revenue in FY2025, together 70.6% of the total — and the filing does not name any of them.

Two movements are worth naming. Customer A’s absolute revenue fell 11.1% in the first half of 2026 (from $158.5m to $141.0m) while Customer B’s rose 41.0% (from $39.9m to $56.3m). And a third 10%-plus customer appeared in FY2025 that had not been there in FY2024. Concentration on the single largest customer, on this measure, has been falling — from 58.2% of half-year revenue to 49.8%.

Which companies A, B and C actually are is a question the filing declines to answer, and we are not going to guess at their identities. What the filing does say, in a different section entirely, is that Samsung Electronics holds 3.77% of Wonik IPS and Samsung Display holds another 3.77%, an identical number of shares each (Source: rcpNo 20260814001845, Note 1-1, as of 30 June 2026). That is a fact about the share register, not about the customer note, and the two should not be welded together without evidence. It is nonetheless unusual enough to state plainly: two of Korea’s largest chip and display manufacturers are on this toolmaker’s cap table.

💰 What Do You Actually Pay for Wonik IPS Today?

At ₩114,800 (as of 4 September 2026) the market capitalisation is $4.19bn, which is 67× FY2025 net profit and 5.8× FY2025 book value — and part of that FY2025 profit was the VM mark.

Against Korean listed peers, on a single consistent date and a single consistent set of FY2025 accounts:

A peer table like this is a snapshot, not a verdict, and every one of these companies is in a different part of the equipment stack: Hanmi builds HBM bonders, HPSP does high-pressure annealing, ISC makes test sockets, and Dongjin Semichem supplies photoresist. They are not substitutes for one another. What the table does establish is that Wonik IPS is not an outlier in either direction on price — the whole Korean tool complex is trading on multiples that assume the current investment cycle continues.

🏷️ Which Wonik Ticker Are You Actually Buying?

There are six listed Wonik-group companies and one delisted one, and the parent holding company trades below the value of its stake in Wonik IPS alone.

🌏 Is Foreign Money Buying or Selling Wonik IPS?

It depends entirely on the window, and one window flips the sign. Over the 20 sessions to 4 September 2026, foreign investors bought a net $112.9m of Wonik IPS (240810). Over the 40 sessions to the same date they are net sellers of $101.5m. Over every session we hold — 55, back to 18 June — they are net buyers of $2.6m.

Read only the 20-session column and the story is “foreigners are accumulating”. Read only the 40-session column and the story is “foreigners are dumping”. Both are arithmetically true of the same stock on the same day. The 40-session column is where the sign flips, and it is the one that shows what the shorter windows are recovering from: heavy foreign selling early in that window, then a buy-back that has, over the full 55 sessions, returned foreigners to roughly flat. Note what the price did meanwhile: across the 40 sessions to 4 September the shares rose 9.4%, from ₩104,900 to ₩114,800, while foreigners sold $101.5m of them. The 31.7% fall people remember belongs to a different window — the 30 June close of ₩168,100 measured to 4 September — and most of it happened before this one opened. Domestic institutions did the mirror image in this one stock — net buyers of $128.1m of Wonik IPS (240810) over the 40 sessions to 4 September, and net sellers of it over every shorter and every longer window we can measure. Neither column is a thesis on its own. This is exactly the trap we set out to avoid in our Foreign Flow Watch series, and a good reminder of how Korean flow data is constructed before you build a thesis on it.

There is one more registered flow worth knowing about, because it is a filing rather than a tape reading. Mirae Asset Global Investments’ declared holding went from 10.01% at 30 June 2026 to 7.31% at 31 July 2026 — a sale of 1,325,386 shares, disclosed on 3 August 2026 as a routine “simple investment” position change (Source: DART 5%-holder report, rcpNo 20260803000366). The same manager’s prior filing put it at 5.19% (2,548,877 shares) as at 17 December 2025 (Source: DART 5%-holder report, rcpNo 20260202001086, which prints that figure as the immediately preceding report), so the position was built and partly unwound inside eight months.

🏛️ Who Controls Wonik IPS, and What Do Minority Holders Get?

Control runs up a three-layer chain — Wonik IPS ← Wonik Holdings (32.90%) ← Wonik Corp ← Horizon LLC (46.33%) — and minority holders get a dividend policy that tops out at a 14% payout ratio in 2027.

Above Wonik IPS, the annual report discloses that Wonik Holdings’ own largest shareholder is Wonik Corp ((주)원익), and that Wonik Corp’s largest shareholder is Horizon LLC (유한회사 호라이즌) with 46.33%. The filing states plainly how that came about: “as a result of a share transfer between related parties on 20 August 2024, the largest shareholder changed. Before: Lee Yong-han (38.18%), Horizon LLC (8.15%). After: Horizon LLC (46.33%).” Lee Yong-han sits on the Wonik IPS board as an inside director. (Source: rcpNo 20260316001453, Section VII.)

Related-party transactions themselves are small: the FY2025 note shows ₩1,387,682 thousand payable to Wonik Holdings and low single-digit-billion-won balances with other affiliates, with no material intra-group sales disclosed.

What did shareholders actually vote on at the 4 September 2026 meeting?

Two items: a new independent director for the audit committee, and — under a brand-new provision of Korea’s Commercial Act — permission for the company to keep its treasury shares instead of cancelling them.

Item 2 is the one we could find no English-language write-up of, because the rule behind it is six months old. On 6 March 2026 Korea promulgated an amendment to the Commercial Act inserting a new Article 341-4 (Obligation to Cancel Treasury Shares). Its first paragraph reads, in the official text: “회사가 자기주식을 취득한 때에는 그 취득한 날부터 1년 이내에 소각하여야 한다” — “when a company acquires its own shares, it shall cancel them within one year of the date of acquisition.” The second paragraph carves out an exception: the company may instead hold them if it prepares a “treasury-share holding and disposal plan” and obtains shareholder-meeting approval, where the purpose is one of five listed cases — the second of which is “using them for employee compensation purposes, such as granting stock options.” Paragraph 3 requires that approval to be renewed every year. Paragraph 4 specifies exactly what the plan must contain. (Source: Korea Ministry of Government Legislation, Commercial Act, Article 341-4, newly inserted by Act No. 21448 of 6 Mar 2026.)

Read Wonik IPS’s proxy statement next to that paragraph and it is a template being filled in. The agenda item lists: (a) purpose — “employee compensation in connection with operating an employee stock-compensation scheme”; (b) type, number and acquisition method — all 428,211 common shares held, acquired by trust contract and other means (merger and split fractional shares); (c) at the holding start date and the planned disposal date — treasury shares 428,211 → 348,211, non-treasury shares 48,655,690 → 48,735,690, treasury ratio 0.9% → 0.7%; (d) expected holding period; (e) expected disposal window, 4 September 2026 to the day before the 2027 annual meeting. Up to 80,000 shares are expected to be delivered to employees before that meeting. (Source: proxy statement filed 20 Aug 2026, rcpNo 20260820000117.)

Two dates sit next to each other, and it is worth being precise about what each one is. The Act’s transitional provision treats treasury shares already held when the law took effect as acquired on “the day six months after the enforcement date” — 6 September 2026. That is the start of the one-year cancellation clock, not a deadline; the deadline it produces is 6 September 2027. Wonik IPS held its meeting on 4 September 2026, two days before that clock started. We are not attributing a motive to the sequencing — the company did not state one, and Paragraph 3 requires the approval to be renewed every year in any case, so the item returns at the 2027 annual meeting regardless. What is on the record is the vote itself: 21.2% of the votes cast did not go to “for” — and the filing combines against, abstentions and anything else into that one figure without separating them. At the 4 September close those 428,211 shares are worth about $36.5m — 0.87% of the company.

For readers tracking Korea’s governance reform, this is the concrete edge of it. The Value-Up program asked companies to cancel treasury stock voluntarily; Article 341-4 now requires it unless shareholders vote otherwise, every year, with an administrative fine of up to ₩50m under Article 635 for the officers the provision names — directors, executive officers, auditors and the like.

This was not the first such vote in Korea, and that is what makes the 21.2% interesting. The rule landed in time for the March 2026 AGM season, and 266 listed companies — 10.7% of the 2,478 December-year-end issuers, 85 on KOSPI and 181 on KOSDAQ — put a treasury-share holding and disposal plan to their shareholders. All 266 were approved (Source: Korea Listed Companies Association and KOSDAQ Association tallies, compiled in Shin & Kim’s 2026 AGM Review No. 2, 20 Apr 2026 — the underlying association survey is not published, so this count rests on that one law-firm tally). Wonik IPS is not an early mover; it is a late one. What separates it from that March cohort is the shape of the vote rather than its timing: it convened a standalone extraordinary meeting in September instead of folding the item into an annual agenda, and more than a fifth of the votes cast did not go to “for” (the filing does not tell us how much of that was outright opposition and how much was abstention). We do not have the against-percentages for the 266 March votes, so we cannot say 21.2% is unusually high — only that it is far from a formality, and that anyone tracking this rule now has a number to compare the next one to.

Who is the new director, and why does his CV matter?

Choi Keun-min spent his career at SK hynix, SK Siltron and SK On — and he is the fourth Wonik IPS outside director drawn from the executive ranks of the company’s own customer base.

Choi Keun-min (born September 1960) joins for a three-year term as an independent director and audit-committee member, taking the board from 6 to 7 directors and the independent share from 50.0% to 57.1% (Source: rcpNo 20260904000161). His career, as the proxy lists it: SK On management advisor (2025–present), SK On CPO and inside director (2024), head of SK On global manufacturing/technology (2023), SK Siltron (2019–2022), SK hynix SHE executive (2018), and head of SK hynix’s production technology centre and technology innovation centre (2017).

He is not the first. The three sitting outside directors are, per the annual report, a former head of Samsung Electronics’ memory manufacturing technology centre and CEO of Samsung Display; a former head of management support at Samsung Electronics’ DS division and at Samsung Electro-Mechanics; and a former head of SK hynix’s future technology research institute. The proxy states the candidate has had no transactions with the company in the last three years. Whether a board substantially staffed by alumni of the customer base is a strength — deep process knowledge, as the board’s own stated rationale argues — or a governance question for minority holders is a judgement each investor makes. It is, at minimum, a fact rarely surfaced in English.

What about the dividend?

Wonik IPS published a three-year dividend policy in November 2025 that raises the payout ratio from about 12% of separate-basis net profit in FY2025 to about 14% in FY2027. The policy is explicit that the ratio is calculated on separate-basis net profit, not consolidated: “a certain proportion of each fiscal year’s net profit on the separate financial statements will be used as the return resource,” rising “approximately 12% (FY2025) / approximately 13% (FY2026) / approximately 14% (FY2027).” (Source: fair disclosure filed 5 Nov 2025, rcpNo 20251105900220.)

For FY2025 the company paid ₩200 per share, ₩9,731m (~$7.2m) in total, against separate-basis net profit of ₩82,565m — a payout of 11.8% on that base, or 11.6% on the consolidated basis as disclosed in the annual report (the filing prints one decimal place). At the 4 September price the yield is 0.17%. This is a company that intends to reinvest, and says so.

📚 Lingo Check

🎯 Why It Matters for K-Export Stars

Nothing in this article is a scoop. Every number is sitting in a public filing on Korea’s DART system, and has been since 14 August. The reason it is not in circulation in English is that the decisive facts are spread across a note on financial instruments, a schedule of investments in other companies, a proxy statement, a name-change announcement from a different company two years ago, and a paragraph of the Commercial Act that was promulgated in March. None of that is translated. A screener will tell you Wonik IPS grew net profit 388%; it will not tell you that a mark on a stock the company owns supplied 69.4% of the pre-tax profit behind it, or that the stock has since halved.

That gap is the whole reason this site exists. Korea’s export champions — the chip supercycle names, their suppliers, their toolmakers — are documented in enormous detail, in a language most of their potential foreign shareholders do not read. The arbitrage is not information nobody has. It is information nobody has read in the right order. If you want to check any of this yourself, the filings are linked throughout, and our guide to buying Korean stocks as a foreign investor and our USD price reference for Korean chip names cover the mechanics.

Currency note: all dollar figures in this article convert Korean won at a single rate, ₩1,345.99/$1 — the rate our own price history records for 4 September 2026, and the single rate used across this whole series of Korean chip-equipment profiles, so that figures in this piece are directly comparable to those in the companion articles. It is applied to figures from earlier fiscal years too, so that every number in the piece is comparable to every other. That is not the rate that applied when those earlier results were earned — the company’s own FY2025 accounts use a year-end rate of ₩1,434.90 per US dollar — so treat the dollar figures for FY2023–FY2025 as a common-yardstick translation, not as reported dollar results. Share prices are shown in won first because that is what you will see on a KRX quote screen.

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 and does not constitute investment advice, a recommendation, or an offer to buy or sell any security. It contains no price target and no rating. Figures are drawn from public regulatory filings and market data as of the dates stated and may be superseded by later disclosures; prices, valuations and shareholdings change. Investing in Korean equities involves risk, including currency risk and the risk of total loss. Do your own research and consider consulting a licensed financial adviser before making any investment decision.