← Deep analysesHome
mttssn research · Nordic Deep Dive
Silex (SILEX.ST)
Semiconductors · MEMS contract foundry (pure-play) · LTM R12 Jul 2025-Jun 2026
Analysis date: 2026-07-28
Price at analysis: SEK 190.15
Method: mttssn_streamlined_v1
Conviction: LOW
HOLD
Conviction: LOW
Silex is a pure-play MEMS contract foundry that listed on Nasdaq Stockholm on 2026-05-07. LTM revenue is SEK 1,510m (+18.6% YoY; Q2 2026 alone +27.3%), adjusted EBIT margin 33.8%, and the balance sheet carries SEK 979m net cash post-IPO. Adjusted ROIC of 27.5% is inflated by undeployed IPO proceeds; a SEK 1.6bn Sweden+US capex build-out through 2030-2034 will pull it lower. Elite operating metrics, thin public history, full valuation.
Adj. ROIC
27.5%
WACC 8% → spread +19.4pp
Economic Profit
+SEK 285M
+SEK 285m EP on SEK 1,467m adjusted IC (27.5% ROIC vs 8% WACC); IC base is temporarily deflated by stripped IPO cash and will rise as the SEK 1.6bn capex program deploys.
FCF Yield
n/a
LTM FCF SEK 199m vs SEK 234m operating cash flow; investment in tangible assets (SEK 232m LTM) already consumes over half of operating cash flow ahead of the announced SEK 1.6bn build-out.
Price / Target
SEK 190 → SEK 195
+3% base; HOLD
Revenue (LTM)
SEK 1.5B
LTM (R12 Jul25-Jun26) net sales SEK 1,510m, +18.6% YoY; Q2 2026 alone +27.3% (+31.8% FX-adjusted), led by development-project design-ins.
EBIT Margin
32.2%
Adjusted EBIT margin 33.8% LTM, up from 28.7% FY2025, tracking the shift toward higher-margin development-project revenue.
EV / IC
n/a
Enterprise value / invested capital
Net Debt
n/a
Net cash SEK 979m post-IPO (net debt/EBITDA -1.6x); ~SEK 961m of that is committed IPO proceeds earmarked for the SEK 1.6bn 2030 capex program, not distributable surplus.
Thesis

Silex is one of a small number of independent, pure-play MEMS contract foundries globally, converting customer design-in projects (51.0% of H1 2026 sales, up from 42.6% a year earlier) into recurring production volume. Growth is broad-based and currently strong: LTM revenue +18.6%, Q2 2026 alone +27.3% (+31.8% FX-adjusted), led by North American telecom/optical-switch and Life Science demand.

The May-2026 IPO (SEK 1bn gross proceeds) and a SEK 750m undrawn RCF fund a defined SEK 1.6bn capacity program through 2030: a ~35% Sweden cleanroom expansion plus, following a binding July-2026 agreement, Silex's first US facility (a converted Onsemi fab in Pennsylvania). Management is explicit the US site will not reach EBIT breakeven before 2029-2030 and will not match the Swedish facility's 2025 economics until 2034 — a long, capital-intensive runway, honestly guided rather than promotionally framed.

Reported quality metrics are elite (adjusted EBIT margin 33.8% LTM, adjusted ROIC 27.5%, net cash), but the record is a streamlined first pass on a company with only 9 quarters of public history, no independently-audited annual report yet on file, and no reverse-DCF computed in the source record — the valuation case below is a rough approximation, not the standard mttssn residual-income panel.

Valuation · reverse-DCF & scenarios

No reverse-DCF exists in the adjustments record (data gap flagged, not filled). A residual-income back-solve (EV - IC = EP / (WACC - g)) on EV SEK 19,906m, IC SEK 1,467m and EP SEK 285m implies the market is pricing roughly 6% perpetual economic-profit growth — plausible given the current 19%+ top-line trajectory, but it leaves no cushion if the capex program slips or margins normalize.

Base SEK 195 (+2.6%) assumes growth normalizes to the high-teens/low-20s% and today's rich ~49x EV/adjusted-NOPAT multiple roughly holds; bull SEK 260 (re-rating as a scarce independent MEMS foundry if the US fab executes ahead of guide); bear SEK 125 (capex-cycle turn, a lost qualification, or US-fab slippage compresses the multiple while capital is mid-deployment).

Market-implied growth
≥26.1%
model ceiling — EV implies more than constant-ROIC sustains
No-growth value / share
SEK 66
35% of price; rest = priced-in growth
ROIC − WACC
+19.4 pp
ROIC 27.5% vs WACC 8.0% — positive = value creation
CAP (priced-in)
8.6 yrs
years of excess returns the price implies (fades to WACC)

At today’s enterprise value the constant-ROIC reverse-DCF maxes out: even at its sustainable-growth ceiling (~26.1%, limited by ROIC 27% ≈ WACC 8%) it cannot reach the current EV. No-growth value is SEK 66/share (35% of price); the market prices in growth and/or a higher ROIC than booked — richly valued on this lens. Read the sensitivity grid.

Scenario24m targetImpl. gUpsideProb.Driver
BullSEK 260≥26%+37%30%US fab conversion executes on/ahead of the 2029-2030 breakeven guide, Sweden +35% capacity ramps smoothly, growth holds 25%+ and margins expand further — market re-rates Silex as a scarce independent MEMS foundry.
BaseSEK 195≥26%+3%45%Growth normalizes toward the high-teens/low-20s%, margins hold near current levels, and the capex program proceeds broadly on the guided multi-year timeline — the current rich multiple roughly holds.
BearSEK 125+21%-34%25%The semiconductor capex cycle turns, a key customer qualification is lost or delayed, or the US fab ramp slips past 2029-2030 — blended ROIC drags as SEK 1.6bn of capital deploys ahead of matching earnings.
Prob-weightedSEK 197+4%100%Scenario-weighted expected value

Sensitivity — fair value / share at WACC × growth

WACC \ g0%3%5%8%10%15%
6.50%8696103114122145
7.25%75828897103121
8.00% (base)6672778489104
8.75%596468747991
9.50%545862677181

Green = fair value above the current price of SEK 190.15. The reverse-DCF conclusion is dominated by WACC and growth — see the whole surface, not a single fair value.

Method & data. NOPAT SEK 403, invested capital and ROIC 27.5% are observed (adjustments.json); WACC 8.0% and terminal g 2.5% are assumptions. EV→equity uses net debt SEK -979. The model holds ROIC constant (no fade) over the explicit horizon; the CAP figure instead fades excess returns to WACC.

⤓ Download the full model (.xlsx) — formula-driven sheets; flex the blue input cells and the model cascades in Excel.

Key drivers

1. Development-project mix shift

Design/NRE revenue rose to 51.0% of H1 2026 sales (vs 42.6% H1 2025) — a leading indicator of future production conversions.

2. North America / telecom-optical demand

North America was 60% of Q2 2026 sales (up from 55% a year earlier), driven by optical-switch demand for AI data centers.

3. Capacity funded by fresh equity

SEK 1bn gross IPO proceeds plus operating cash fund a SEK 1.6bn Sweden+US build-out through 2030.

4. Margin leverage on fixed cleanroom capacity

Adjusted EBIT margin rose to 33.8% LTM from 28.7% FY2025 as higher-margin development revenue outgrows fixed costs.

5. First US manufacturing facility

The Onsemi Pennsylvania fab acquisition (signed 2026-07-07) shortens lead times to Silex's largest customer base and reduces geopolitical/tariff exposure.

Key risks
Conclusion

Silex is a genuinely high-quality, fast-growing MEMS foundry, but this is a first-pass streamlined read of a company with only 9 quarters of public history, and the SEK 20.9bn market cap already prices in durable ~19%+ growth and margin resilience through a SEK 1.6bn, multi-year US+Sweden capex program.

HOLD, low conviction: no margin of safety at SEK 190, base target SEK 195 (+2.6%). Revisit on the FY2025 Annual Report (customer concentration, audited multi-year history) or a pullback toward the SEK 125-150 bear range.

Footnote evidence & sources

Every adjustment traces to the cited note/page in the source filing — click to open it there. Latest interim: 📄 open

Adjustment / figureValueSourceWhy mttssn treats it this way
LTM/R12 net sales1,510Condensed consolidated income statement, R12 column 📄 p.11Company-computed rolling-12m (Jul 2025-Jun 2026) column used directly as the LTM revenue anchor; no manual quarter roll-forward needed.
LTM/R12 EBIT (reported)487Condensed consolidated income statement, R12 column 📄 p.11Reported operating profit before mttssn adjustments; matches the company's own R12 EBIT figure.
Non-recurring items add-back (IPO/ownership-change consulting costs)23Note 9, Non-recurring items table, R12 column 📄 p.21Consulting costs tied to the IPO and the July-2025 change of ownership; genuinely one-off, added back pre-tax consistent with the company's own Adjusted EBIT definition — drives our EBIT bridge to exactly match the company's Adjusted EBIT (0% divergence).
Intangible assets / goodwill0Condensed consolidated balance sheet, all 3 periods 📄 p.12Zero intangible assets and zero goodwill disclosed at 2026-06-30, 2025-06-30 and 2025-12-31 — no capitalized R&D or M&A-driven goodwill exists to reverse, atypical for the SOFTWARE sector template; Silex is a PP&E-heavy manufacturing foundry.
Total equity (IC anchor)2,586Condensed consolidated balance sheet, 2026-06-30 📄 p.12Latest Q-snapshot equity used as the IC base per streamlined methodology (BS values from latest Q, not FY-anchor).
Accumulated OCI reserve32Balance sheet equity section, 'Reserves' line 📄 p.12Equity 'Reserves' line represents accumulated translation + cash-flow-hedge reserve (per statement of comprehensive income, H1 2026 CFH fair-value change -16, translation +1); stripped from equity per OCI methodology.
Interest-bearing debt188Note 9, 'Capital employed' reconciliation, 2026-06-30 column 📄 p.21Liabilities to credit institutions, current (18) + non-current (170) = 188, cross-checked against the company's own Capital employed build.
Lease liabilities (tracked, excluded from IC)138Balance sheet: current (59) + non-current (79) lease liabilities 📄 p.12ROU assets (342) are 9.8% of total assets / 24.6% of non-current assets — not the company's PRIMARY operating asset (owned buildings/machinery/CIP dominate at 1,006) — so lease_liabilities_in_ic=false per the decision rule; IFRS16 lease interest already sits below EBIT regardless, so no NOPAT add-back either way.
Cash and cash equivalents1,305Balance sheet, 2026-06-30 📄 p.12Includes ~SEKm 961 net IPO proceeds received May 2026; only ~2% of LTM revenue (~SEKm 30) treated as operational, remainder (~SEKm 1,275) stripped as excess cash per methodology — see qualitative_flags on why this 'excess' is actually committed growth capital.
Effective tax rate0.21Income statement, R12: tax -101 / PBT 480 📄 p.11R12 effective tax rate used to tax-effect adjusted EBIT into adjusted NOPAT.
Company Adjusted EBIT (APM)510Note 9, Adjusted EBIT calculation, R12 column 📄 p.21Company's own APM bridge — matches our adjusted EBIT exactly, confirming a clean bridge with no undisclosed adjustments.
FY2025 net sales / EBIT (baseline)1,385Condensed consolidated income statement, Jan-Dec 2025 column 📄 p.11Full FY2025 comparative disclosed directly inside the Q2 2026 interim report — used as the FY anchor baseline in lieu of a standalone annual report; the prospectus fallback was not needed.
Subsequent event — US fab acquisition40Significant events after the end of the period 📄 p.2Binding agreement dated 2026-07-07 (after period-end) to acquire a Pennsylvania semiconductor fab from Onsemi for USDm 40; not yet in the 2026-06-30 balance sheet but material to forward capital deployment — flagged qualitatively, not modelled into current-period IC.
Quality · Buffett tenets7 / 15
Understandable business
Pure-play MEMS contract foundry: revenue splits development-project (design/NRE) and production work, a legible model — but only 9 quarters of public disclosure exist since the 2026-05-07 IPO, so the track record is short.
Durable moat
[byteskostnader · vidgas] development-project revenue rose to 51.0% of H1 2026 sales (vs 42.6% H1 2025), and ROCE was positive in all 9 disclosed quarters (14.9%-22.5%, always above the 8% WACC); emerging — no audited multi-year record yet; falsifierare: a rival foundry (X-FAB/Tower/Teledyne) displaces Silex in a qualified design, or the ~60%-North-America revenue concentration is hit by reshoring/tariff shifts.
Management & capital allocation
[allokering · candor] net IPO proceeds (~SEK 961m) plus operating cash are earmarked for a numbered SEK 1.6bn Sweden+US capex program; our independently rebuilt adjusted EBIT (SEK 510m R12) matches the company's own APM exactly (0% divergence), and management candidly guides the US fab to no EBIT breakeven before 2029-2030; röd flagga: the SEK 1.6bn program is pre-revenue-committed with a long payback — execution slippage is the observable that would cut this score.
Financial strength & returns
Adjusted ROIC 27.5% vs 8% WACC (+19.5pp spread, EP +SEK 285m); net cash SEK 979m, 74% equity ratio; ROCE stayed positive across all 9 disclosed quarters. Caveat: ~SEK 1,275m of the IC base is un-deployed IPO cash — un-stripped ROIC is only 13.9%, and both figures should mean-revert as the capex program lands.
Valuation margin of safety
No reverse-DCF block exists in the source record (streamlined first pass); a residual-income back-solve (EV-IC = EP/(WACC-g)) implies the SEK 20.9bn market cap already requires ~6% perpetual EP growth, and EV/adjusted-NOPAT runs ~49x — the price sits at-to-above a conservative fair value with no discount to cushion execution risk.