← back to picks

HQ Zhunan · Taiwan Reporting TWD Credit synth Aaa/AAA · net cash Affiliate Hon Hai (Foxconn) Valuation 2026-05-31 Damodaran FCFF · 1000-iter MC FCFF · Dark v3

A near-pure terminal bet on a single customer · +178% base headline, but β-contingent

Base MoS says BUY — revised to HOLD: 77.5% of value is terminal, no balance-sheet floor

Base intrinsic value NT$889/share (after a 10% governance haircut; pre-haircut NT$988) vs market NT$320. The headline holds only if you underwrite both the suppressed-β pin (1.10) and intact ~16% margins. Strip the perpetuity and the floor is ~NT$208 — below today's price. Net cash is just 2.7% of value; this is a single-customer (Applied Materials) terminal bet, not a balance-sheet bargain.

SectorSemiconductor equipmentCountry mixUS 62% · TW 22% · KR 13%β / MC σ1.10 pinned · ±NT$85/sh (1000 runs)GovernanceHon Hai control · 10% haircutQualityNet cash · synth AAA · never a loss since 2015 IPOTerminal load77.5% of value is perpetuity · no floor

Intrinsic / share

NT$889.08

post 10% gov · pre NT$987.87

Market / share

NT$320.00

late-May 2026 · TWSE

Margin of safety

+177.8%

vs intrinsic

Enterprise value

NT$105.56B

81.2% terminal

Cost of equity / debt

6.25% / 1.22%

β 1.10 · CRP 0.42%

Terminal ROIC / g

10.00% / 1.13%

spread ~423bp (ROIC 10.00% vs WACC 5.77%)

What it sells, where it sells

Operating segments

Precision modules for WFEEtch / deposition subsystems — #1-ranked Applied Materials supplier; build-to-print, single-anchor-customer order book~80%

Inspection / metrologyNew growth engine — revenue +150% YoY in 4Q25, ~10% of Q4 revenue; management sees potential to double again in 2026~20%

Foxsemicon does not disclose a clean product split — the business is one order book driven by global wafer-fab-equipment (WFE) capex, overwhelmingly through the Applied Materials relationship. The metrology line (+150% YoY in 4Q25) is the only diversifier away from single-customer module work; the April-2026 Frantron (富蘭登) bolt-on adds aerospace MRO. Neither is yet large enough to break the AMAT dependence that makes 77.5% of the DCF a single-customer perpetuity.

Country mix (revenue-weighted CRP input)

🇺🇸United States62%

🇹🇼Taiwan22%

🇰🇷South Korea13%

🇯🇵Japan3%

Weighted by end-customer / capex geography, not production-site location. US 62% reflects AMAT (anchor) plus Intel and Micron; Taiwan 22% is TSMC capex and the home listing; Korea 13% is the Samsung / SK Hynix memory complex. ~60% of production physically sits in China (Kunshan / Songjiang) serving international customers — that asset/expropriation risk is carried in the 10% governance haircut, not the CRP mix, which lands at a low ~0.42%.

Background — five things to know before the case

Context the share price doesn't carry on its face. Skim this once and the bull / bear bullets below stop reading as inside baseball.

The revised verdict is HOLD — β-contingent, not a clean BUY. The base case prints +178% MoS, but a concentration-aware re-examination found the buy survives the terminal-ROC sweep and either single stress alone, yet not the compounded cell single-customer risk actually produces: AMAT operating-margin compression and the sector-implied cost of capital, acting together. That cross-term lands NT$271–288 — below the NT$320 quote.
77.5% of value is the discounted terminal block; net cash is just 2.7%. The engine's own decomposition: PV of explicit FCFF 18.3%, net-cash bridge 2.7%, PV of terminal value 77.5%. Strip the perpetuity and the "floor" is ~NT$208/share at the 5.77% base WACC (NT$174 at 9%) — both below the NT$320 quote. Net cash alone is ~NT$27/share. There is no balance-sheet floor under this buy.
The β 1.10 pin is the valuation. The 5Y regression β (~0.80) is plausibly suppressed by thin float plus the Hon Hai overhang; Damodaran's global Semiconductor-Equipment β_u (2.117, 391 firms) re-levers to ~2.30 at the company's D/E, implying WACC 9–10% and a fair-to-overvalued stock on base assumptions. The council pinned a sector-aware compromise of 1.10 — above the suppressed regression, well below the ASML/AMAT-heavy aggregate. The single most important number in a 78%-terminal DCF is an unfalsified analyst choice.
#1-ranked Applied Materials supplier into the AI-WFE up-cycle. 2025 revenue grew 26.7% (second straight year >26%) to NT$20.84bn; operating profit +16.8% to NT$3.11bn, both record highs. AMAT raised its 2026 semicap revenue-growth guide to >30%; order visibility now runs to end-2026 and management says it is "being chased by orders." Metrology revenue rose ~150% YoY in 4Q25.
Thai-plant ramp is depressing the FY25 margin — transitory, but load-bearing for the thesis. Two new Thai plants (Rayong opened Jan-2025, Chonburi Nov-2025) at ~60% utilization push 4Q25 gross margin to 23.9% and FY25 operating margin to 14.9%. The base case recovers margin toward 16% by year 10 (mid-point of the 12–20% historical band). If AMAT's single-buyer leverage instead compresses margin below ~13%, the cross-term breaks the buy.
Hon Hai (Foxconn) group affiliate — control with very low economic ownership. Control runs through Hon Yang VC (~6.29% direct); total board+insider holding only ~6.7%. The May-2026 AGM lifted directors' non-compete restrictions, letting directors sit on parallel Hon Hai-group boards — a genuine minority-shareholder agency overhang, offset by a clean capital-allocation record (consistent dividend (NT$11 FY25, NT$28–39 in prior years), no dilution, no tunneling evidence). Nets to the 10% governance haircut.

Where we diverge from the sector

Each row asks: if we'd used the sector median for this one assumption instead of our override, how much would the share price change? Negative = our override is more conservative (worth less); positive = more aggressive (worth more). Skim down the rightmost column to see which bets are doing the work.

Assumption

Our input

Sector median

Value impact

01Revenue growth10-year revenue CAGR vs Semiconductor-Equip median

~8.5% CAGR13% Y1-5 (below recent 26% spikes) fading to 1.13% terminal — AMAT content gains + metrology ramp + Thailand

~6.0% CAGRSemiconductor Equip, global cross-sector median

+NT$95

02Operating marginYear-10 target vs 2025 depressed base

16.0%Mid-point of the 12-20% historical band; recovers from the 14.9% Thai-ramp trough — the load-bearing operating lever

~12%Semiconductor Equip sector median operating margin

+NT$140

03Sales-to-capitalReinvestment efficiency vs sector median

1.50×Company's through-cycle revenue/invested-capital (~1.53 at base) — held both phases

~1.5×Semiconductor Equip industry standard

NT$0

04Terminal ROCPerpetual excess return vs engine default

10.0%Faded from the engine's ~19% — a single-customer captive cannot hold a 14pt perpetual excess return; sweep ±7% only

~19%Engine-resolved (16% margin × 1.5 S2C × 0.79)

−NT$70

05Cost of capital10y WACC vs WACC implied by sector-relevered β

5.77%β 1.10 pinned (compromise over the suppressed ~0.80 regression) · CRP 0.42% · D/V low single digits — the single load-bearing input

~9-10%Damodaran Semi-Equip β_u 2.117 re-levered → β 2.30 → fair-to-overvalued

+NT$480

Net effect of overrides

The +178% headline is dominated by the β pin (row 5): re-levering to the sector β_u (2.30) lifts WACC to 9–10% and collapses the perpetuity that is 77.5% of value. Terminal ROC alone is value-inelastic (±7%). The real levers are the discount rate (β) and the operating margin — and the cross-term that compounds both lands NT$271–288, below the NT$320 quote.

+NT$569

Our override Sector median Override adds value Override subtracts value

The story & the five claims

The 10-year story this DCF is built on, plus the five anchor claims that translate the story into engine inputs.

The 10-year story

Foxsemicon is the #1-ranked Applied Materials precision-module supplier riding the AI-driven wafer-fab-equipment up-cycle — net cash, never a loss since its 2015 IPO, and a Hon Hai (Foxconn) affiliate. The base case prints a +178% margin of safety, and on its own assumptions it is genuinely cheap into a structural AI capex super-cycle. But the headline is not the fortress it looks like. This is, arithmetically, a near-pure terminal bet: 77.5% of value sits in the perpetuity, net cash is only 2.7%, and stripping the terminal block leaves a floor around NT$208 — below today's NT$320 price. The whole valuation hangs on two underwriting calls. First, a β of 1.10, hand-pinned over a suppressed ~0.80 regression because re-levering Damodaran's sector β_u (2.117) to 2.30 would push WACC to 9–10% and make the stock fair-to-overvalued. Second, an operating margin that recovers from the 14.9% Thai-ramp trough back toward 16% rather than being ground down by Applied Materials' single-buyer transfer-pricing leverage. Terminal-ROC durability — the moat question everyone debates — barely moves the answer (±7%). The real levers are the discount rate and the margin, and the cross-term that compounds an AMAT operating shock with the sector-implied WACC lands NT$271–288, fifteen percent below the quote. So the +178% is real only in the pinned-β, intact-margin world. That is why the verdict is HOLD, not BUY: the margin of safety exists, but it rests entirely on the two contingencies, with no balance-sheet floor to catch a fall.

Two debates worth pressure-testing

Is the β really 1.10, or is the suppressed regression hiding a 2.30 sector reality?

Our view: Genuinely unresolved — and it is the valuation. The 5Y regression (~0.80) is plausibly suppressed by thin float and the Hon Hai overhang; the council pinned 1.10 as a compromise. But Damodaran's Semi-Equip β_u 2.117 re-levers to 2.30 at this D/E → WACC 9–10% → fair-to-overvalued. In a 78%-terminal DCF the single most load-bearing number is an unfalsified analyst choice. If you take the sector β at face value, the BUY is gone.

Does the terminal-moat (ROC) question decide the verdict?

Our view: No — and the original note used it as a straw man. The ROC sweep (0.07 ↔ 0.13) moves value only ±~7%, because it froze WACC and β. The load-bearing input is the denominator (WACC − g). The real risks are AMAT-driven margin compression and a sector-aware cost of capital — and compounded, they land NT$271–288, below the NT$320 quote.

CLAIM 01Revenue compounds ~13% Y1-5, fading to the 1.13% TWD risk-free.growth_high: 13% · terminal: 1.13% · NT$20.8B → NT$51.0BAbove the 10-16% WFE-market growth on AMAT content gains, metrology ramp (+150% YoY 4Q25), Thailand capacity, and the Frantron bolt-on — but below the recent 26% spikes to respect cyclicality and the high 2025 base.

CLAIM 02Operating margin recovers from the 14.9% Thai trough toward 16%.target_op_margin: 16% by Y5 (Thai plants mature)Mid-point of the 12-20% historical band as the Thai Megasite reaches scale and mix shifts to higher-margin metrology / repair-spares / aerospace. The load-bearing operating call: below ~13% the buy breaks.

CLAIM 03Reinvestment at the company's through-cycle 1.50× sales-to-capital.S2C: 1.50× Y1-5 and Y6-10Matches the ~1.53 realized at base. Capital absorbs the Thai Megasite build-out; no efficiency stretch assumed in either phase.

CLAIM 04Terminal ROC faded to 10% — not the engine's ~19% perpetual moat.override_roc: 10% · terminal_g: 1.13%A single-customer (AMAT) captive supplier cannot defend a 14pt excess return forever. 10% keeps a healthy WACC premium for switching costs while refusing an unfaded perpetual-moat assumption. Value-inelastic — sweep ±7%.

CLAIM 05No failure risk — net cash, never a loss; but no balance-sheet floor either.failure: 0% · gov haircut: 10% · net cash ~NT$27/shSynth Aaa/AAA, ICR ~49×, never a loss since the 2015 IPO. But net cash is only 2.7% of value — strip the 77.5% terminal block and the floor (~NT$208) sits below the NT$320 price. 10% haircut for Hon Hai control + non-compete waiver.

Where we diverge from sell-side

β pinned at 1.10, not re-levered to the sector 2.30 — this single call is the headline. The regression (~0.80) is plausibly thin-float-suppressed; re-levering Damodaran's Semi-Equip β_u (2.117) to the company's D/E gives 2.30, WACC 9–10%, and a fair-to-overvalued stock. The +178% lives entirely in the gap between those two betas. Worth ~NT$480/share in the ladder.
Terminal ROC faded to 10%, not the engine's ~19%. A captive single-customer supplier cannot hold a 14pt perpetual excess return against a ~6% WACC. The fade costs ~NT$70 — but, crucially, the ROC question is not where the buy breaks (sweep ±7%); it was used as a straw man in the original note.
Margin recovers to 16%, not held at the 14.9% Thai trough. Mid-point of the 12-20% band as the Thai plants mature. If AMAT's single-buyer leverage instead compresses margin below ~13%, the operating leg of the cross-term fires.
Governance haircut 10%, above the 5% public-company baseline. Hon Hai control with ~6.7% economic ownership plus the AGM's blanket non-compete waiver is a real minority-agency overhang; the China production footprint's asset risk is carried here, not in the CRP mix.
The cross-term the base run never compounded is the whole story. AMAT operating shock (g→8%, margin→13%) + the stress regime's 8.88% WACC + 15% governance lands NT$271–288 — below the NT$320 quote. The additive-stress framing (each stress alone clears) flattered the name to a false "every scenario +53%" fortress.

Two-sided case — bear anchors

The β is really ~2.3, not 1.10. If the suppressed regression is an artifact and the sector relever is the truth, WACC is 9–10%, not 5.77%. On base operating assumptions that alone makes the stock fair-to-overvalued — and the buy breaks at a reachable WACC ~10.8% even before any margin haircut.
AMAT compresses margin via single-buyer transfer pricing. The real single-customer risk is not "AMAT stops buying" — it is structural margin compression. If margin grinds to ~13% instead of recovering to 16%, the buy breaks at a far more reachable WACC ~7.9%, and the compounded cross-term lands NT$271–288, ~10-15% below today's price.
No balance-sheet floor. Strip the 77.5% terminal block and the floor is ~NT$208 at base WACC (NT$174 at 9%) — both below NT$320. Net cash alone is ~NT$27/share. There is nothing under this buy if the perpetuity assumptions fail; it is a near-pure terminal bet.
Cyclical WFE roll + high 2025 base. 2025's +26.7% partly reflects customer pre-stocking; management concedes 2026 must "fill" that base. A WFE down-leg plus the Thai depreciation drag could keep the optical margin near the 14.9% trough longer than the model assumes.
China / Hon Hai geopolitical and agency tails. ~60% of production sits in China serving international customers (asset/expropriation overhang); the non-compete waiver lets directors run parallel Hon Hai-group businesses. Captured at 10% governance today, but a step-up here is the kind of re-rating the thin float would amplify.

Risks to thesis (tail, not bear case)

β is really the sector 2.30High

The single load-bearing input. If the suppressed ~0.80 regression is wrong and Damodaran's Semi-Equip relever (2.30) is right, WACC is 9-10% not 5.77% — fair-to-overvalued on base assumptions. No scenario stress-tested β; the buy breaks at WACC ~10.8%.

AMAT margin compressionHigh

Single-buyer transfer-pricing leverage grinds margin below the 16% target toward ~13%. Compounded with a sector-aware WACC the cross-term lands NT$271-288 — below the NT$320 quote. On shocked cashflows the buy breaks at WACC ~7.9%.

No balance-sheet floorMed

77.5% of value is terminal, net cash only 2.7%. Strip the perpetuity and the floor is ~NT$208 (NT$174 at 9% WACC) — below price. There is nothing to catch a fall if the terminal assumptions fail.

WFE cycle roll / high 2025 baseMed

2025 +26.7% partly reflects pre-stocking; 2026 must "fill" the base. A capex down-leg plus Thai depreciation keeps the optical margin near the 14.9% trough longer than modeled.

China production + Hon Hai agencyLow

~60% of production in China serving international customers (asset/expropriation overhang); AGM non-compete waiver lets directors run parallel group businesses. Priced via the 10% governance haircut; a step-up is a tail.

TWD strength on reported numbersLow

2025 reported EPS fell ~10% YoY largely on TWD-appreciation FX losses despite record revenue. Optical, not operating — but it is what the market anchors on, and a strong TWD dents reported net again.

10-year forecast

Revenue NT$23.6B → NT$51.0B over 10y (13% Y1-5 on AMAT content gains + metrology ramp + Thailand, fading to the 1.13% TWD risk-free). Operating margin recovers from the 14.9% Thai-ramp trough to 16% by Y5 (mid-point of the 12-20% historical band) and holds — the load-bearing operating assumption.

Monte Carlo distribution

The Monte Carlo spreads the base case (p5 NT$743, all 1,000 draws above NT$320) — but it varies inputs around the pinned β of 1.10 and intact ~16% margins. It does not sample the sector-relevered β (2.30) or the compounded AMAT operating shock. The real downside lives in that un-sampled cross-term (NT$271–288), which is why the verdict is HOLD, not the +178% the distribution implies.

Mean NT$880.58 ± NT$85.33/sh, 1000 iterations (0 failed). P(intrinsic < market NT$320.00) = 0.0%.

⚠ Active diagnostic: terminal_growth (0.0113) >= risk_free_rate (0.0113); Damodaran's stable-growth ceiling is the risk-free rate

Cost of capital build

Risk-free rate	1.13%
Mature-market ERP	4.23%
Levered β	1.10
Weighted CRP	0.42%
Cost of equity	6.25%
Pre-tax cost of debt (synth Aaa/AAA)	1.53%
D / V	~9%
WACC	5.77%

Full year-by-year DCF

Year	Revenue	Op mgn	EBIT	EBIT(1−t)	Reinvest	FCFF	PV
1	NT$23.55B	15.14%	NT$3.57B	NT$2.82B	NT$1.81B	NT$1.01B	NT$956M
2	NT$26.61B	15.35%	NT$4.09B	NT$3.23B	NT$2.04B	NT$1.19B	NT$1.06B
3	NT$30.07B	15.57%	NT$4.68B	NT$3.70B	NT$2.31B	NT$1.39B	NT$1.18B
4	NT$33.98B	15.78%	NT$5.36B	NT$4.24B	NT$2.61B	NT$1.63B	NT$1.30B
5	NT$38.40B	16.00%	NT$6.14B	NT$4.85B	NT$2.95B	NT$1.91B	NT$1.44B
6	NT$42.48B	16.00%	NT$6.80B	NT$5.38B	NT$2.72B	NT$2.66B	NT$1.90B
7	NT$45.99B	16.00%	NT$7.36B	NT$5.84B	NT$2.34B	NT$3.51B	NT$2.38B
8	NT$48.69B	16.00%	NT$7.79B	NT$6.20B	NT$1.80B	NT$4.40B	NT$2.83B
9	NT$50.40B	16.00%	NT$8.06B	NT$6.43B	NT$1.14B	NT$5.30B	NT$3.24B
10	NT$50.96B	16.00%	NT$8.15B	NT$6.52B	NT$380M	NT$6.14B	NT$3.57B

Methodology & flags

Damodaran FCFF DCF, 10y explicit + perpetuity, in TWD. Risk-free 1.13% (local-currency government bond). Synthetic credit Aaa/AAA. CRP 0.42% from revenue-weighted country mix × Damodaran 2026 CRPs. Terminal ROIC faded to 10.00%; 10% governance haircut applied post-DCF. Monte Carlo: 1000 iterations. Engine v1.0.0 · result: /Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/foxsemicon/output/2026-05-31-result.json

R&D cap: OFF · Lease cap: OFF · Failure: OFF · ESO: OFF
Governance haircut: 10% applied post-DCF (NT$987.87 > NT$889.08)
Sensitivity tornado: not run

FCFF DCF · dark theme v3. Generated 2026-05-31T20:06:52 from /Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/foxsemicon/output/2026-05-31-result.json.