Halfway through 2026, the same question keeps coming back across the semiconductor rally: how much of the price is cash flow already delivered, and how much is option premium not yet earned? This piece looks at a company that pushes the question to its extreme — a European supplier where almost all the good news is already booked, and almost all of it is already in the price.
VAT Group (VACN.SW on the SIX Swiss Exchange, hereafter VAT) makes vacuum valves for semiconductor process equipment, and it is the global hidden champion of this layer: the commonly cited market figure puts its global share at about 70% — note that this number has never appeared in the company’s own filings, a point we return to below. It does not sell to chipmakers; its customers are the equipment makers. The pick-and-shovel seller to the pick-and-shovel sellers. On the quality side: EBITDA margins ran between 30.0% and 31.2% across FY2023 to FY2025, including FY2023, when revenue fell about 23% and the margin still held above 30%; FY2025 revenue reached CHF 1.074 billion, with record free cash flow of CHF 230 million. On the cycle side: Q1 2026 orders came in at CHF 356.3 million, up 47% year over year — the second-largest quarter in the company’s history — with a book-to-bill of 1.6x and a backlog up 42% from the end of 2025. For VAT, the WFE (wafer fab equipment) recovery is not a forecast. It is signed orders in hand.
The market has paid for both sides at once: a trailing PE of about 84x, an FCF yield of about 1.3%, and a share price around CHF 625 — less than 1% below the 52-week high. Sell-side targets are a lagging and divided snapshot: the older consensus range aggregated by data providers, CHF 510 to 607, tops out below the current price, while UBS raised its target to CHF 745 on June 10 — the “fully priced” judgment here rests on the multiple and the cash-flow yield themselves, not on price targets.
In Leeno: The Chip-Testing Company Taking Its Own Quality-Premium Test, we ran the same hidden-champion framework on Korea’s test-probe maker: that was a quality premium with a governance discount already charged, once. VAT is the mirror image — cycle, quality, and scarcity premiums priced simultaneously, with no discount of any kind. The question of this piece: what is left to buy at 84x?
In April 2026, VAT published its Q1 2026 trading update. The same release carried two sets of numbers pointing in opposite directions.
On the order side: Q1 orders of CHF 356.3 million, up 47% year over year (up 67% at constant currency), the second-largest quarter in the company’s history; a book-to-bill of 1.6x; and a backlog of CHF 431.3 million, up 42% from CHF 304 million at the end of 2025. On the sales side: Q1 sales of CHF 220.9 million, down 20% year over year. Management’s explanation: logistics disruptions from the Middle East conflict, with an estimated impact of CHF 20 to 25 million, expected to be recovered in Q2. A rough decomposition: the gap between the nominal and constant-currency figures, about 11 percentage points, is currency; the logistics impact equals roughly 7 to 9 percentage points of the prior-year base — strip out both, and quarterly sales were broadly flat.
The market chose to price the orders, not the sales. Over the past year, the stock has climbed from a 52-week low of CHF 257.50 to about CHF 625 in early June 2026 (a closing-price reference), a gain of more than double across the range, leaving it less than 1% below the 52-week high of CHF 629.60. UBS had upgraded the stock from “Neutral” to “Buy” back on November 20, 2025 (with a price target of CHF 380 at the time), citing the order recovery and improving FCF; on June 10, 2026, it kept the “Buy” and raised the target to CHF 745 (as reported by Investing.com; relayed here purely as market information). Separately, alongside the March annual results, the company proposed raising the dividend 12% to CHF 7 per share.
The orders are real. The rally is real. The problem sits in the third number: 84x.
The critical steps of semiconductor manufacturing — etch, deposition, EUV lithography — all happen inside vacuum chambers. Valves handle chamber isolation, opening and closing, and pressure control: wafers pass through valves on the way in and out, process gases switch through valves, and the vacuum environments of adjacent chambers are kept apart by valves. In a tool’s bill of materials the valve is not the big-ticket item, but it is the gate that decides whether the tool runs stably. That is what “component chokepoint” means here: low unit price, no way around it.
Advanced nodes are structurally increasing valve content, for three reasons (synthesizing company IR and analyst-report framings). First, multi-patterning: even in the EUV era, chipmaking still requires repeated etch-plus-deposition loops, each step needing vacuum isolation, so the valve count per advanced etch tool rises as nodes progress. Second, new process steps: the GAA architecture at the 2-nanometer node introduces entirely new processes such as cryogenic etch and ALD, and every new recipe is a new chamber-isolation requirement — a new step directly equals new valves. Third, High-NA EUV demands more precise vacuum management, where VAT holds a design-binding advantage: 132 spec wins in 2024 and another 150 across FY2025, up 14% (company annual-report basis). A spec win means the valve is written into the tool’s design specification — switching suppliers means changing the design.
The leverage is in the weighting: etch and deposition together account for roughly 45% of WFE spending, and VAT’s share of valves for these tool categories is commonly cited at about 70%. Advanced nodes increase the spending weight of etch and deposition, so VAT benefits over-proportionally from a WFE recovery.
Here we must stop for this piece’s first sourcing caveat. The figure of about 70% comes entirely from secondary citations by analysts and media: buy-side reports commonly cite a range of 65% to 75%, and Morningstar’s 2025 report carries the title “Dominant Vacuum Franchise” — but VAT’s own annual reports and press releases have never given a precise percentage; the official language is only “leading global supplier.” Wherever this piece uses the number, it is labeled “commonly cited market figure.” Another frequently quoted number — VAT valves in 70% to 80% of installed High-NA EUV tools — likewise rests on secondary sources only, with even less support, and this piece does not rely on it.
On revenue structure, FY2025 group revenue of CHF 1.074 billion splits three ways. Semiconductor valves: CHF 725 million, about 67%. Advanced Industrials (display, solar, and other industrial vacuum applications): CHF 150 million, about 14% — a segment that has shrunk continuously from CHF 194 million in FY2023, down 23% over three years. Global Service (spare parts, repairs, retrofits): CHF 199 million, about 19%, up 19% year over year, with more than 90% of it coming from the semiconductor market (company IR basis). The company does not use the “recurring” label, but the service business follows fab utilization rather than capex, making it the least cyclical layer in the mix.
The margin’s resilience has been tested through a downturn: in FY2023, revenue fell about 23% and net profit about 38%, yet the EBITDA margin held at 30.6%; FY2024 came in at 31.2%; FY2025 at 30.0%, with the strong franc a main cause of the dip. For reference, the company’s historical peak, FY2022, saw revenue of roughly CHF 1.14 billion and an EBITDA margin of about 35% — note these are approximations back-derived from the FY2023 decline, not verified against original filings in this round. Against Leeno’s four straight years of operating margins above 42% (different basis: that is an operating margin, this is an EBITDA margin, and directionally Leeno’s is higher), VAT sits a tier lower — it is, after all, tied to equipment, with one-off orders and only the roughly 19% service layer as a cushion. The quality is real, but it is not top of class.
Lay out the order sequence. FY2023 orders: CHF 692 million (the down-year). FY2024: a jump to CHF 1.033 billion. FY2025: flat at CHF 1.033 billion (up 6% at constant currency) — within which semiconductor valve orders of CHF 676 million fell 5.3%, still digesting the aftershock of customer destocking, while Global Service orders of CHF 212 million rose 21.1%. The backlog stood at CHF 304 million at the end of FY2025, down 18% year over year (from CHF 370 million at the end of FY2024) — up to that point, the backlog was still the tail of the downturn.
Then comes the Q1 2026 step-change: quarterly orders of CHF 356.3 million — more than a third of FY2025’s full-year total in a single quarter; a book-to-bill of 1.6x; a backlog up 42% in one quarter to CHF 431.3 million — the inflection happened inside that single quarter. Management issued full-year guidance at the same time: 2026 orders, revenue, EBITDA, EBITDA margin, net profit, and FCF are all expected to exceed 2025; Q2 sales are guided to CHF 265 to 295 million; the macro anchor is industry research forecasting 2026 WFE at about USD 130 billion, up roughly 10% year over year, with the company expecting to benefit over-proportionally in vacuum-intensive segments.
This is the strongest card in the bull hand: for most semiconductor supply-chain companies, “the WFE recovery” is still a forecast; for VAT, it has already become contracts. A 1.6x book-to-bill means the delivery ramp in Q2 and Q3 is, with high probability, mechanical.
On Stock Analysis figures, the trailing PE is about 84x (EPS of CHF 7.14), against a market value of about CHF 17.8 billion (as of May 18, 2026), implying an FCF yield of about 1.3%. One basis note: computed directly off the early-June price of about CHF 625, the multiple is higher still (about 87x, with a yield of about 1.2%) — the snapshots are two to three weeks apart; this piece quotes the former throughout, which is the conservative end.
A second basis issue matters more than the snapshot: the denominator. The 84x sits on FY2025 trailing earnings — net profit of CHF 214 million, a pre-inflection number. Cyclical stocks carry naturally inflated trailing PEs at an inflection point: the numerator is already pricing the recovery while the denominator is still parked in the down-year. Given Q1 orders up 47%, the Q2 guidance, and the mechanical delivery ramp described above, that denominator will rise mechanically and the multiple will compress with it — yet even with the earnings recovery loaded into the denominator, the multiple likely only falls from the mid-eighties to a visibly lower tier, and the order-of-magnitude judgment of “expensive” stands. The point still has to be laid on the table: without it, “fully priced” gets dismissed with a single “that’s trailing.” The same applies to the roughly 1.3% FCF yield: its numerator is pre-inflection FY2025 cash flow (CHF 230 million), so the yield will lift as the recovery lands — again not by enough to rewrite “expensive.”
Three things are packed into the 84x. The first layer is the cycle premium: orders in hand, a 1.6x book-to-bill, full-year guidance raised — the highest-certainty layer, but also the most transparent, transparent to the point of leaving no expectation gap. The second layer is the quality premium: three years of EBITDA margins above 30%, an FCF margin of about 21%, record free cash flow of CHF 230 million — proven repeatedly in the accounts. The third layer is the scarcity premium: a commonly cited share of about 70%, plus the design binding of spec wins, makes this one of the few “unavoidable” exposures to a WFE recovery.
What deserves closer attention is what 84x does not buy. It does not buy share expansion — if the 70% figure is accurate, the ceiling is already overhead, and future growth can only equal market growth. It does not buy a margin step-up — the FY2025 margin was still being eroded by the franc (down 1.2 percentage points year over year), with Malaysian capacity only a partial hedge. And it does not buy any discount — no governance discount, no cycle discount, no customer-concentration discount. Full price.
The sell-side targets need a basis note here: the consensus rating across 16 analysts is “Buy,” and the average price targets aggregated by data providers range from CHF 510 to 607 — even the top of the range sits below the current price — but that range is a lagging snapshot with wide dispersion: on June 10, 2026, UBS raised its target to CHF 745, above the current price (MarketScreener / Stock Analysis / Investing.com basis; all of the above is relayed purely as market information and does not represent this piece’s view). Price targets are therefore not reliable evidence of either overshoot or undervaluation; this piece’s “full pricing” judgment rests solely on the 84x PE and the roughly 1.3% FCF yield themselves.
The UBS note, though, deserves a head-on fight; neutral relaying is not enough — it is the strongest single card on the other side. Its argument goes a step beyond a simple bull call: consensus earnings forecasts are too low and carry room for upward revision (the headline itself reads “upside risk to consensus”). That is a direct rebuttal of “fully priced”: if the denominator still has systematic room to be revised up, then “full” is premature — the line itself has not been drawn yet. This piece’s response: that is precisely the structure laid bare. The current multiple holds only as long as the act of revising earnings upward keeps repeating; the moment upgrades stop — even without downgrades — one layer of support is pulled away. That is not a rebuttal of zero tolerance for error. It is another way of writing it.
This forms an exact mirror with Leeno. Leeno: operating margins in the 45% range, a trailing PE of about 52x, plus a governance discount on the order of 15% already charged, in one stroke — what the market bought was “the discount has been taken.” VAT: margins a tier lower (30% on an EBITDA basis), a multiple about 60% higher, and no discount ever charged. Two fairness qualifiers belong inside the mirror. First, the gap between 52x and 84x is not all discount layers — there is also a growth differential: VAT holds contracted order momentum (Q1 up 47%, book-to-bill of 1.6x), while Leeno has no in-hand order evidence of comparable magnitude. Second, zero discount does not equal mispricing: Leeno’s discount corresponds to a governance event that actually happened, while VAT has had no event at all — “no discount” may simply be “no bad news.” With the qualifiers in place, the mirror still holds. The same hidden-champion framework, two opposite answers: Leeno’s question is whether the discount taken was enough; VAT’s question is why no discount needs to be taken at all.
First, the cycle cuts both ways, and the demonstration is recent. In FY2023, with WFE in decline, VAT’s revenue fell about 23% and net profit about 38% — chokepoint status does not exempt the cycle, only the margin. The current 84x prices in WFE growth of +10% arriving on schedule; if AI capex decelerates or leading fabs push out expansions, orders and the valuation would contract together. From a starting point of 84x, multiple compression does not require earnings to turn negative.
Second, customer concentration has numbers but no names — this piece’s second sourcing caveat. VAT’s 2025 annual report discloses: the three largest customers account for roughly 16%, 14%, and 13% of FY2025 revenue, about 43% combined (about 45% in FY2024) — all anonymous, with no top-ten figure disclosed. The only customer that public documents can name is Lam Research (which gave VAT a Supplier Excellence Award in 2023); whether Applied Materials, Tokyo Electron, or ASML sit among the top three cannot be confirmed from public filings. The risk of losing a single customer therefore cannot be precisely assessed — only confirmed to exist structurally: nearly half of revenue hangs on three anonymous buyers.
Third, a 70% share is both the moat and the ceiling. In industrial products, a share like this leaves almost no room for further expansion; growth is entirely a bet on the total WFE pie. The longer-term structural risk: if new computing routes (quantum, photonics) reduce dependence on conventional vacuum-chamber processing, the core barrier gets bypassed — not a near-term variable, but at 84x the duration assumption has to carry it.
Fourth, the franc. VAT reports in CHF, its costs sit in Switzerland and Europe, and its revenue is largely dollar-denominated, so a strong franc compresses margins directly — the FY2025 EBITDA margin fell 1.2 percentage points, with currency one of the causes the company explicitly named. At an 84x valuation, a one-point margin drift is not noise.
Fifth, Advanced Industrials keeps shrinking. With display and solar demand weak, the segment has contracted from CHF 194 million in FY2023 to CHF 150 million in FY2025, about 14% of the group — a corner of the growth story that keeps leaking air.
Sixth, the explanation for the Q1 sales decline still rests on management’s account. Logistics disruption of CHF 20 to 25 million, recovered in Q2 — a claim that is verifiable but not yet verified, and the test is whether Q2 sales land inside the CHF 265 to 295 million guidance range. If Q2 misses, the “it was logistics, not demand” narrative gets re-examined.
The public data behind this piece has seven gaps; the first three bear directly on judgment. First, no primary source for the share figure: the roughly 70% global share has never appeared in VAT’s official disclosures, and this piece labels it “commonly cited market figure” throughout — if the true share sits at the bottom of the cited range (65%), the scarcity-premium story loses some force. Second, the customer list is invisible: the identities of the top three customers are not publicly disclosed, with only Lam Research indirectly confirmable via the supplier award, so customer-structure risk can be quantified no further than “top three at about 43% combined.” Third, the historical valuation band is missing: we could not find VAT’s own five-year PE range, nor a systematic comparison against peers such as MKS Instruments or Entegris — how expensive 84x is relative to the company’s own history, this piece can only answer in absolute terms (an FCF yield of about 1.3%), not in relative terms.
The remaining four: FY2022 peak financials are back-derived approximations (revenue of roughly CHF 1.14 billion, not verified against original filings); the China business carries only management’s qualitative “continues to grow,” with no absolute value or share disclosed; FY2024 net profit of CHF 212 million is inferred from the wording of the FY2025 annual report and has not been independently checked against a line item; the 70% to 80% High-NA EUV valve share rests on secondary sources only and is not used in this piece. One further note: the share price here is an early-June 2026 closing-basis reference, not a precise single-day close, and the PE and market-cap snapshots sit two to three weeks apart from the price snapshot.
This piece therefore has the same standing as its mirror: a structural observation on how the market prices the component layer of the supply chain, not a conclusion about a single stock. Its core judgment is this — the order book has already proven the cycle, the accounts have already proven the quality, and the market has paid for cycle, quality, and scarcity in full, holding back no discount of any kind. Full pricing is not necessarily wrong pricing; it simply hands over the entire margin for error: 84x is a trailing starting point, and the earnings recovery will mechanically compress the multiple, but what remains after the compression is the same zero-tolerance structure — results do not need to get worse, they only need to stop getting better. The first test sits in the Q2 delivery numbers.
This report is an independent KSINQ market observation for informational purposes only. It is not investment advice. Price data are an early-June 2026 snapshot (closing basis, not a precise single-day value); financial and order data follow the company’s IR disclosures. Written June 13, 2026.