Siltronic AG
Capital-intensive merchant wafer manufacturing sold to chipmakers largely under multi-year long-term agreements (LTAs) with customer prepayments; pricing outside LTAs is spot-cyclical; 30.0M shares outstanding, anchor shareholder Wacker Chemie (~24% after its May 2026 secondary placement, down from 30.8%)
Earnings, margins, COGS & capex
FY2025 (results release 2026-03-12): sales EUR 1,346.7M (-4.7%), EBITDA EUR 316.9M (23.5% margin), but a swing to net loss (EUR -77.9M, EPS EUR -2.31) as depreciation from the ~EUR 2B FabNext Singapore 300mm fab ramps ahead of matching revenue, plus USD weakness, spot-price pressure outside LTAs, and the shutdown of the small-diameter (SD) line (roughly one third of the sales decline). Capex fell to EUR 369.1M from EUR 523.4M; net financial debt ended 2025 at EUR 836.5M. No dividend for 2025. Q1 2026 (results release 2026-04-29): sales EUR 306.5M (-17.5% q/q), EBITDA EUR 65.1M (21.2% margin), EBIT EUR -52.4M, net result EUR -66.8M (EPS EUR -1.92), free cash flow EUR -89.2M, net debt EUR 935.5M - in line with guidance. In May 2026 Wacker Chemie placed 2.1M existing shares (7% of capital) at EUR 89.35 via accelerated bookbuild, cutting its stake to ~24% - a secondary sell-down with ~EUR 188M gross proceeds to Wacker, none to Siltronic.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~77¢ is cost of goods and ~24¢ operating expense, and the remainder is an operating loss.
Revenue trend
Margins
down; FY2026 guided 20-24%
negative through 2026 as FabNext depreciation (EUR 490-520M guided) front-runs revenue; FY2026 EBIT guided significantly below prior year
loss-making; recovery hinges on volume/price upcycle
COGS structure
Dominated by polysilicon feedstock (historically sourced heavily from anchor shareholder Wacker Chemie), energy (a structural cost disadvantage at German sites Burghausen/Freiberg), depreciation on fab equipment, and labor. Depreciation is the swing factor: guided EUR 490-520M for FY2026, exceeding guided EBITDA at the midpoint - the accounting core of the current loss phase.
Capex
EUR 369.1M FY2025, EUR 523.4M FY2024 - the tail of the ~EUR 2B FabNext 300mm greenfield fab in Singapore (first wafers 2024, ramping with customer demand). FY2026 guided EUR 180-220M, the lowest in years, pivoting the story from build-out to harvest/utilization. Q1 2026 net capex EUR 110.4M was front-loaded within that guide.
Latest earnings
In line - company titled the release 'Siltronic closes Q1 2026 in line with expectations'; no formal consensus beat/miss disclosed
FY2026 confirmed: sales mid-single-digit % below FY2025 (roughly flat adjusted for FX and SD closure, at EUR/USD 1.18), EBITDA margin 20-24%, EBIT significantly below prior year, capex EUR 180-220M, depreciation EUR 490-520M, net cash flow around prior-year level (FY2025: EUR -85.3M)
- EBITDA margin (Q1 2026)
- 21.2% (vs 23.3% Q4 2025)
- Net financial debt (Mar 2026)
- EUR 935.5M (vs EUR 836.5M at year-end 2025)
- FY2026 capex guide
- EUR 180-220M (vs EUR 369.1M FY2025)
- FY2026 depreciation guide
- EUR 490-520M (exceeds guided EBITDA at the midpoint)
Growth drivers
- AI-driven 300mm wafer demand — the FY2025 release notes AI 'is shifting demand noticeably toward 300 mm wafers'; Q1 2026 call cited AI end-market growth continuing to strengthen
- FabNext Singapore ramp - modern 300mm capacity positioned for the upcycle without new peak capex
- LTA book with prepayments providing a revenue floor and visibility through the downcycle
- Normalization of customer inventories (300mm relatively robust — 200mm power-segment inventories still persistently high)
- Capex step-down to EUR 180-220M enabling a swing back toward positive free cash flow after 2026
Bull & bear
A leveraged play on the AI wafer upcycle at the trough of its own investment cycle: the ~EUR 2B FabNext build is done, capex halves in 2026, 300mm demand is inflecting on AI logic and HBM memory, and any price/volume recovery drops through a depreciation-heavy but cash-light cost base into rapid FCF and earnings normalization - with strategic-interest optionality on top.
- Capex falls from EUR 523.4M (2024) to EUR 180-220M (2026 guide) exactly as AI-driven 300mm demand strengthens - the classic cycle setup where cash flow inflects before earnings
- EBITDA is real (EUR 316.9M, 23.5% margin in a trough year); the net loss is depreciation timing, not runaway cash burn - net cash flow guided around EUR -85M and set to turn as capex falls
- Memory-customer capacity constraints plus relatively robust 300mm demand point to wafer tightness into 2027; Siltronic has fresh, qualified Singapore capacity to sell into it
- Wacker's May 2026 placement of a 7% block (2.1M shares at EUR 89.35) was absorbed by institutional demand near market - free float rises, and the shrinking anchor stake keeps strategic outcomes live
- Consolidation reference point: GlobalWafers offered EUR 145/share (~EUR 4.35B) in 2021/22; the deal lapsed on regulatory timing, not on price or antitrust refusal
- At EUR ~2.5B market cap the stock trades near ~1.9x trough sales (EV ~2.6x) with a rebuilt asset base that would cost far more to replicate today
A price-taking, USD-exposed, energy-cost-disadvantaged #4 player whose stock roughly doubled year-on-year on AI hope while still guiding declining sales, sub-depreciation EBITDA, negative FCF, and EUR ~0.9B net debt - if the 300mm upturn slips into 2027-28 or Chinese capacity caps pricing, the equity re-rates back down hard.
- FY2026 guidance is a decline: sales mid-single-digit % lower, EBITDA margin 20-24% (down from 23.5%), EBIT significantly negative, net cash flow still ~EUR -85M - the fundamentals have not yet turned, only the narrative has
- The stock is up ~98% y/y and roughly tripled off its 52-week low (EUR 31.70) to above EUR 100 in June 2026 - while Q1 sales fell 17.5% q/q; expectations now front-run delivery by several quarters
- Depreciation of EUR 490-520M/yr against guided EBITDA of roughly EUR 260-310M (20-24% margin on declining sales) means accounting losses through at least 2026-27, deferring dividends and screening poorly on earnings
- FX asymmetry: revenue in USD, costs in EUR/SGD; every cent of dollar weakness below the 1.18 planning rate cuts guidance with no natural hedge
- 200mm power-segment inventories are 'persistently high' per the company itself, and the SD-line closure permanently removed revenue; Chinese state-backed 300mm entrants are a structural price ceiling on the mid-range
- Wacker Chemie has begun selling down (30.8% to ~24% in May 2026) - a persistent ~24% overhang - and the placement raised money for Wacker, not Siltronic; a longer downturn could force a genuinely dilutive raise
What it is worth
EV/EBITDA and EV/Sales vs wafer peers (Shin-Etsu, SUMCO, GlobalWafers) plus reverse-DCF sanity check; all figures vintage Jul 2026
EUR 40-55/share
upcycle slips to 2028, dollar stays weak, 200mm glut persists, Chinese capacity caps pricing; EBITDA stuck near ~EUR 280M with EUR ~0.9B net debt and dilution risk from a forced raise
EUR 80-100/share
gradual recovery, FY2027 EBITDA EUR 380-450M, FCF turning positive, ~8x EV/EBITDA
EUR 120-145/share
2027 upcycle with FabNext ramping, EBITDA EUR 500M+, ~8-9x EV/EBITDA, plus strategic-bid optionality
At EUR ~2.5B market cap (30.0M shares x EUR ~84) + EUR ~0.94B net debt (Mar 2026), EV ~EUR 3.4-3.5B = ~2.6x FY2025 sales and ~11x trough FY2025 EBITDA (EUR 316.9M) - a full multiple for a loss-making trough year, cheap only if mid-cycle EBITDA returns: Siltronic earned EUR 433.9M EBITDA (28.7% margin) in 2023 and materially more in the 2022 record year, so EUR 450-550M by 2027-28 is the recovery case, not fantasy. The current price effectively pre-pays for the 300mm upcycle arriving on schedule; GlobalWafers' lapsed EUR 145/share offer (2021/22, ~EUR 4.35B) is a dated but real strategic reference point. Scenario bands below are illustrative multiples math, not price targets. Not financial advice.
SWOT
Strengths
- Top-4 global merchant silicon-wafer position (with Shin-Etsu, SUMCO, GlobalWafers) in an effectively oligopolistic, high-barrier market
- Leading-edge 300mm epitaxial and polished wafer technology qualified at top logic and memory customers - qualification cycles of 1-2 years lock in share
- LTA contract book with customer prepayments cushions downcycles and part-funded FabNext
- New, efficient FabNext Singapore capacity already built - upcycle leverage without new peak capex
- Equity ratio still 42.4% (Mar 2026) despite the downcycle and heavy investment phase
Weaknesses
- Loss-making: FY2025 net loss EUR -77.9M, Q1 2026 EPS EUR -1.92; depreciation wave (EUR ~500M/yr guided) exceeds current EBITDA run-rate
- Net financial debt EUR 935.5M (Mar 2026) against EUR ~2.5B market cap and still-negative free cash flow; no dividend for 2025
- USD revenue vs largely EUR/SGD cost base — a weak dollar (guidance rests on 1.18) directly compresses sales and margin
- High German energy and labor costs at legacy Burghausen/Freiberg sites vs Asian peers
- Subscale vs Shin-Etsu and SUMCO in R&D and capacity share — price-taker outside LTAs; 200mm power-segment exposure stuck in an inventory glut
Opportunities
- AI/HBM-driven 300mm wafer-area growth — AI is shifting demand noticeably toward 300mm (company commentary), with memory-customer capacity constraints a classic precursor to wafer tightness and price recovery
- FabNext utilization ramp converting sunk capex into incremental EBITDA at high drop-through
- LTA repricing in the next tight market; capex holiday enabling an FCF inflection from 2027
- Potential strategic/consolidation interest — GlobalWafers' EUR 145/share takeover offer (~EUR 4.35B) lapsed in early 2022 only because German regulatory approval did not arrive before the offer deadline; Wacker's ongoing stake reduction keeps the register open
- SiC/GaN and advanced substrate adjacencies over time
Threats
- Prolonged 200mm inventory overhang in power/industrial and delayed memory recovery
- Structural Chinese wafer capacity build-out (state-backed 200mm/300mm entrants) eroding mid-range pricing over time
- Further USD weakness below the 1.18 planning rate directly cuts guidance
- Trade/export-control disruptions in semiconductor supply chains — concentration among a few large chipmaker customers
- Balance-sheet strain if the downcycle extends — an equity raise at depressed prices would dilute (none has occurred to date; the May 2026 placement was a Wacker secondary, not new capital)
Moats, dependencies & bottlenecks
Moats
Wafer requalification at logic/memory fabs takes 12-24 months and risks yield; switching suppliers mid-node is rare - incumbency at top-tier logic and memory customers is sticky
The top 4-5 merchant suppliers control the large majority of the 300mm market; greenfield entry costs EUR ~2B+ and years of qualification - though Chinese state-backed entrants chip at the low end
Decades of accumulated defect-density and flatness know-how; hard to replicate, but Shin-Etsu and SUMCO are at least peers
Multi-year volume/price agreements smooth cycles and co-fund capacity, but expire and reprice with the cycle
Roughly #4 by share with high-cost German legacy sites; FabNext Singapore improves but does not close the gap to Shin-Etsu
Dependencies
Hemlock Semiconductor, Tokuyama) Hyperpure electronic-grade polysilicon comes from few qualified sources; Wacker is both a key supplier and ~24% shareholder - alignment today, concentration risk structurally
Samsung, Intel, Infineon, SK Hynix, Micron class) Revenue concentrated among a handful of large logic/memory/power makers; their capex and inventory cycles set Siltronic's demand
Sales priced in USD, costs in EUR/SGD; FY2026 guidance explicitly rests on EUR/USD 1.18 - dollar weakness was a named driver of the 2025 decline
Energy-intensive crystal pulling at Burghausen/Freiberg is exposed to European power costs
The AI thesis transmits to wafers mainly through memory and leading-edge logic capacity additions; company cites capacity constraints at memory customers as a current headwind
Advantages
- Only European-headquartered top-tier 300mm wafer maker — supply-chain-diversification appeal for Western chipmakers and EU/US resilience agendas
- Freshly built, modern Singapore 300mm capacity while peers' next expansion decisions still lie ahead
- Prepayment-backed LTA model reduces downcycle cash risk versus pure spot exposure
- 42.4% equity ratio preserved through the peak-capex phase - leveraged but not distressed
Weaknesses
- Loss-making with negative free cash flow through the guidance horizon; no dividend for 2025
- Price-taker outside LTAs in a market where Shin-Etsu and SUMCO set the pace
- Structural FX mismatch (USD revenue, EUR/SGD costs)
- High-cost German legacy footprint; SD line already shuttered
- ~24% Wacker stake in active sell-down mode (7% block sold May 2026) - persistent share overhang
Bottlenecks
- Depreciation wave from FabNext (EUR 490-520M/yr guided) exceeding EBITDA - earnings cannot turn positive until volume/price recovers
- 200mm power-segment customer inventories — persistently high per Q1 2026 commentary, gating a full-portfolio recovery
- FabNext utilization ramp is customer-demand-paced; empty cleanroom space earns nothing while depreciating
- Balance-sheet headroom: EUR ~0.9B net debt and negative FCF limit strategic flexibility until cash flow turns
Top signals & trends
Top signals
AI shifting demand noticeably toward 300mm per company commentary; the single most important upcycle confirmation
Capacity constraints at memory customers cited in Q1 2026 - new fabs mean wafer demand with a lag
Still 'persistently high' as of Q1 2026; the lagging half of the portfolio
Dollar weakness directly cuts guidance; watch each quarter's realized rate
May 2026: sold a 7% block (2.1M shares at EUR 89.35, ~EUR 188M gross to Wacker), cutting its stake from 30.8% to ~24%; further sell-downs could pressure the stock or open the register to a strategic buyer
Each qualification converts depreciation drag into contribution margin
The swing to positive free cash flow (2027 on current guidance trajectory) is the pivotal fundamental milestone
Trends
The core demand thesis; wafers are the unavoidable substrate of the AI capex boom
Western chipmakers value a European-headquartered qualified wafer source
State-backed 200mm/300mm entrants pressure mid-range pricing structurally (competitive-analysis context only, not an investment view on any Chinese name)
The source of the 200mm inventory glut; recovery timing uncertain
Peers also deferred expansions; supply growth into the next upcycle is constrained, supporting price
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Electronic-grade polysilicon supplier AND ~24% anchor shareholder (post May-2026 secondary placement, down from 30.8%)
US polysilicon producer - alternative hyperpure feedstock source
Japanese electronic-grade polysilicon supplier
German crystal-growing systems and metrology equipment maker
World's largest foundry; leading-edge logic wafer demand
Memory + foundry wafer consumer
IDM logic customer
HBM/DRAM leader - key AI-memory transmission channel for wafer demand
US memory maker; HBM/DRAM capacity additions
Power semiconductor customer - the 200mm inventory-glut segment
European analog/power customer
Specialty foundry customer
Analog IDM; 300mm analog capacity build-out
Global #1 silicon wafer maker; superior scale, margins, and balance sheet
Global #2; pure-play wafer peer facing the same cycle with bigger 300mm share
Taiwan-listed #3, the only wafer maker with fabs across Europe, Asia, and the US (incl. new Texas 300mm); its EUR 145/share Siltronic takeover offer lapsed in early 2022 on German approval timing
Captive-plus-merchant supplier, strong in Korea with SK Hynix proximity
Chinese state-backed capacity building at 200mm/300mm; structural pricing pressure at the mid/low end (competitive context only)