NYSE:TSM Taiwan Semiconductor Manufacturing Company Limited

ISIN: US8740391003
TechnologySemiconductorsAdvanced-Node Foundry
NYSE ADR (1 ADR = 5 TWSE ord.) · Hsinchu, Taiwan · Mkt cap ~$2.09T · Beta ~1.37 Analysis Status: On-Going
All figures USD (NYSE ADR); company reports in TWD. Mode-B refresh; prior report 2026-06-16 at $425.83. Q2 2026 reported 15-16 Jul — this report factors the post-earnings tape.
$402.30
+1.0%
20 Jul 2026 · Signal v6

Changes Since Last Report vs. 2026-06-16 (HOLD/BUY/STRONG BUY @ $425.83)

All three signals unchanged: HOLD / BUY / STRONG BUY. The story is the post-earnings tape: TSMC reported a record Q2 (rev +36%, GM 67.7%, EPS +77%, FY guide raised >40%) on 15-16 Jul, yet the stock fell ~4.6% on the $60-64B capex raise + AI-bubble fears. Net: cheaper price + higher earnings → Valuation improved, Timing deteriorated. Price −5.5% ($425.83 → $402.30).

DISCLAIMER: This is a quantitative framework for educational purposes only. It is not financial advice. Always do your own research and consult a licensed financial advisor before making investment decisions.

Taiwan Semiconductor Manufacturing Company Limited

TSMC is the world's largest dedicated semiconductor foundry — it manufactures chips designed by others (Nvidia, Apple, AMD, Broadcom, Qualcomm) rather than selling its own branded products. Its economic engine is leading-edge logic: it holds an estimated 90%+ share of the most advanced process nodes (3-nanometre today, 2-nanometre ramping), the only place at scale where the fastest AI accelerators and smartphone processors can physically be made. That near-monopoly on the cutting edge, married to industry-best manufacturing yields and CoWoS/SoIC advanced-packaging capacity, is its moat: customers design around TSMC's process and cannot easily move. It reports in New Taiwan dollars but is analysed here as the USD NYSE ADR (1 ADR = 5 Taiwan ordinary shares). Think of it as the indispensable toll-booth of the AI and computing supply chain.

HorizonSignalComposite ScoreConfidenceKey Driver
Short-term (1–3 mo)HOLD5555%Base BUY-accumulate capped to HOLD by the short technical-confirmation cap — post-earnings tape rolled over (daily downtrend, below the 50-DMA, bearish confluence); Technical & Catalyst entry groups both unmet. Buy on confirmation.
Medium-term (6–12 mo)BUY6662%Dominant Quality (88) + Valuation now Attractive after the de-rate (forward P/E ~21x, 49% Street upside) carry the matrix; Neutral medium macro pressure blocks STRONG BUY.
Long-term (3–5 yr)STRONG BUY7465%Wide-moat AI-toll-booth foundry + Strong Tailwind AI driver (88) + long-horizon Tech/EM macro Tailwind (XLK long O) amplify the base BUY to STRONG BUY.
Next update: 2026-07-31 — AI mega-cap prints (22-30 Jul — Alphabet/Meta/etc., the AI-capex read-through that just moved TSM) + FOMC 29 Jul + Q2 GDP 30 Jul cluster; refresh the trading day after (31 Jul). Own Q3 print ~mid-Oct.
Table of Contents
1Five-Pillar Scorecard2Hard Gates & Do-Not-Buy Status3Pillar Detail: Business Quality4Pillar Detail: Valuation Attractiveness5Pillar Detail: Underlying Drivers6Pillar Detail: Economic Alignment7Pillar Detail: Entry/Exit Timing8Economic Event Risk9Multi-Timeframe Technical Analysis10Price Chart (6-Month Daily)11Scenario Summary12Entry / Exit Rules13Position Sizing Context14Calibration Snapshot15Data Sources & Methodology
1

Five-Pillar Scorecard

Five independent scores — each 0–100 with its own confidence. The three fundamental pillars (Quality / Valuation / Timing) set the base BUY/HOLD/SELL via the Decision Matrix; the two context pillars (Underlying Drivers, Economic Alignment) then amplify a BUY to STRONG BUY or a SELL to STRONG SELL when both corroborate.

Business Quality

88
elite — >90% advanced-node share, 67.7% GM
conf 82%

Valuation Attractiveness

64
attractive on fwd earnings; rich on FCF/book
conf 78%

Entry/Exit Timing

49
neutral — post-earnings tape rolled over
conf 55%

Underlying Drivers

88
Strong Tailwind (AI/HPC)
conf 72%

Economic Alignment

55
Contrarian (S) / Trend-Following (L)
conf 68%
2

Hard Gates & Do-Not-Buy Status

Binary safety checks — any TRIGGERED gate is a hard cap regardless of the scores above; CAUTION gates are sizing notes.
Financial Distress
Net cash (debt/equity 0.17, cash/share ~$652 ord-equiv), interest coverage 176x, current ratio 2.49. Nowhere near distress.
Earnings Event Risk
Q2 2026 already reported 15-16 Jul; next print (Q3) is ~mid-Oct, well outside the 14-day window. No earnings blackout.
⚠️
Valuation Ceiling
Forward P/E ~21x is below the warranted 28x and the Semis 28x guardrail → ceiling does NOT fire. Caution only: trailing P/E ~30-31x sits right at the 28x guardrail line and P/B 10.2x / P/S 14.7x are rich; on forward (cycle-appropriate) earnings the name is not expensive. Price $402 is far below the $700 high and $600 median targets.
Accounting / Dilution
Share count flat (~5.19B ADR-equiv, no dilution), minimal SBC, clean audit. Q2 non-operating income ~14% of net (interest on the cash pile + a one-off FX/investment gain) is below the 30% distortion line and below the 15% normalisation trigger — earnings are clean operating earnings, NOT AI-mark-up-inflated.
⚠️
Regulatory / Binary (Taiwan)
The Taiwan/China geopolitical tail is a persistent, structural risk (a conflict or blockade would be an existential, >20% event) — but it is NOT a dated, pending binary ruling, so Gate 5 is flagged CAUTION, not fired. Carried in the Bear scenario and position-sizing, not as an automatic cap.
3

Pillar Detail: Business Quality

A deep dive into the Quality score: business economics, moat, ROIC and the industry benchmark.
Business Quality — Pillar Score
Elite foundry monopoly on the leading edge; record 67.7% GM, 55% net margin, net cash
88
confidence 82%

Lifecycle & sector: Technology · Semiconductors (advanced-node foundry). Classified Growth (dominant, cash-generative leader) — Q2'26 revenue +36% YoY yet ~55% quarterly net margin and net cash. Scored on the Semiconductors profile: gross margin, capacity utilisation, ROIC, cycle context, moat. Q2 2026 (reported 15-16 Jul) strengthened the picture — record gross margin, guidance raised — so Quality holds at 88.

Sub-signalTSMCRead vs peers / cycleScore
Revenue trajectoryQ2'26 US$40.2B, +36% YoY / +12% QoQ (high end of guidance); FY26 guide raised to >40% growthExtraordinary for a $2T-cap; AI/HPC now the majority driver; 7nm-and-below = 77% of wafer revenue94
Profitability / marginGross margin 67.7% (record, up from 66.2%); operating margin ~60%; net margin ~55% (qtr)Far above the >60% "strong" semis bar; pricing power on leading-edge wafers is intact and expanding95
Cash generationOperating-cash-flow margin ~59%; but FCF depressed — capex raised to $60-64B (P/FCF ~59x, FCF yield ~1.7%)The one soft spot: the AI build-out is consuming FCF now (a growth choice, not weakness), but it is the reason the stock sold off62
Balance-sheet healthNet cash; debt/equity 0.17; interest coverage 176x; current ratio 2.49Fortress — self-funds the $100B Arizona + $60B+ annual capex without stress92
Capacity utilisation (industry benchmark)Advanced-node (3nm/5nm) utilisation >90%; 2nm rampingDemand-constrained at the leading edge, not supply-glutted — the healthy side of the semis cycle92
Industry benchmark — Gross Margin + Capacity Utilisation. GM 67.7% (record) + advanced-node utilisation >90%. Both well past the "GM >55% + Util >80%" top band. Benchmark score: 93/100. Cycle note: semis are cyclical, so this is a strong-cycle read — but AI/HPC demand visibility (CEO guided robust through 2029-2030) makes a near-term trough less likely than in prior PC/mobile cycles.
Cost advantage
92
Structural scale + best-in-industry yields at the leading edge; no competitor can match cost-per-good-wafer on 3nm/2nm.
Switching costs
85
Customers co-design to TSMC's PDK/process; re-porting a chip to another foundry costs years and yield risk. Trimmed slightly for Intel 18A wins (see below).
Pricing power
90
Repeated leading-edge wafer price increases absorbed by the market; GM expansion to 67.7% is the proof.
Intangible assets
85
Process IP, 65k+ engineers, decades of manufacturing know-how; a genuine technology lead of 1-2 nodes.
Network effects
60
Ecosystem/OIP design-enablement flywheel (more customers → more IP blocks → more customers) — real but weaker than a two-sided marketplace.

Moat average ≈ 82 — one of the widest moats in the analysed universe; the competition-exposed dimensions (Switching Costs, Cost Advantage) are scored from the named-rival read below, not in the abstract.

Competitive Environment

TSMC is gaining-to-stable at the leading edge, holding an estimated >90% share of advanced logic. But two credible challengers are worth naming, and one wrinkle emerged this month. Overall competitive threat: moderate (nudged up from low as Intel 18A lands its first external foundry wins).
RivalThreat typeShare trajectory (TSMC vs rival)Moat-erosion vector
Intel Foundry (18A)Direct advanced-node foundry rivalTSMC dominant, but Intel 18A is newly winning external deals (reported Apple & Tesla foundry agreements; stock +278% H1'26 on the turnaround)The one genuine new erosion vector — a credible second leading-edge source would trim switching costs & pricing power over time. Still early / yield-unproven at scale
Samsung FoundryDirect advanced-node foundry rivalTSMC extending its lead; Samsung sub-scale on 3nm, persistent yield gapLimited — Samsung has not closed the yield/customer gap; low near-term threat
Hyperscaler custom silicon (Google TPU, Amazon, Meta MTIA)Substitution for merchant GPUsNeutral-to-positive for TSMC — these custom chips are still fabbed by TSMCShifts demand between TSMC's customers (Nvidia vs in-house), not away from TSMC; a demand-mix issue, not share loss

Net effect on the moat: → Switching Costs held at 85 and Cost Advantage 92 (Intel 18A is the reason they are not 90+/95). This propagates to the §11 Bear trigger (Intel/Samsung advanced-node share gain) and the §12 thesis-invalidation rule (a credible second leading-edge source at scale).

ROIC & capital allocation (86/100). ROE ~36%, ROA ~17%, ROIC well above cost of capital through the cycle. Disciplined allocation: reinvests at high returns (the $60-64B capex + $100B Arizona is the AI build-out), a growing dividend (~0.9% yield, ~26% payout), no value-destructive M&A. The near-term FCF compression is a deliberate growth investment, and the balance sheet funds it comfortably.
4

Pillar Detail: Valuation Attractiveness

Sector-appropriate multiples, FCF yield, reverse-DCF implied growth, embedded optionality, and the analyst-consensus cross-check.
Valuation Attractiveness — Pillar Score
Attractive on forward earnings after the post-earnings de-rate; rich on FCF/book
64
confidence 78%
Warranted-multiple anchor (the 40% weight). Semiconductors, high-quality (BQ 88): discount rate r = 4.55% 10-Y (FRED, 17 Jul) + 4.5% ERP + 0.0% risk add-on = 9.05%; disciplined growth g_near 15% (secular-growth sector cap; consensus multi-year EPS CAGR is ~30%, haircut and capped), g_term 3%. Two-stage warranted P/E ≈ 28.1x, capped at the Semis 28x guardrail28x. Actual forward P/E ~21x (NTM) after the de-rate → ratio 21/28 = 0.75 → Attractive. The offsets that pull the composite below the anchor: trailing P/E ~30-31x (Fair-to-rich, right at the guardrail line), P/B 10.2x, P/S 14.7x (rich), and a 1.7% FCF yield (expensive on cash, capex-driven). Net Valuation 64 — Attractive on the earnings the cycle is actually delivering, tempered by the cash/book richness.
Earnings-quality decomposition (step 7b) — clean. Q2 non-operating income was ~14% of net income (interest income on the large cash pile plus a one-off ~NT$69B FX/investment gain, ~10% of net). That is below the 15% normalisation trigger and far below the 30% distortion gate; TTM is even lower (prior quarters carried negative non-operating items). TSMC is explicitly NOT in the S&P-500-concentration / non-operating-gains AI cohort — it is a foreign foundry earning clean operating margins from selling wafers, not a US index mega-cap booking mark-ups on private-AI stakes. Multiples are scored on genuine operating earnings; no material adjustment needed (clean_pe ~ reported).
MultipleTSMCReference readScore
Forward P/E (NTM)~21x; ~19x on FY27 EPSBelow warranted 28x and below the stock's own strong-cycle history; cheap vs AI mega-cap peers (ASML ~30x)74
Trailing P/E~30-31xAt the 28x Semis guardrail line — Fair-to-rich, but trailing understates given +77% EPS growth50
PEG (fwd)~1.05 (FMP fwd); 0.59 trailingReasonable-to-cheap against ~30% forward EPS growth70
P/B · P/S10.2x · 14.7xRich in absolute terms — the capital-intensive richness the FCF yield also flags38
FCF yield (universal anchor): ~1.7% (P/FCF ~59x). In the 1-3% "expensive" band — the honest counterweight to the cheap forward P/E. It is depressed by the $60-64B growth capex, not by weak operations (operating-cash-flow margin ~59%). A reader is paying up on cash today for the AI build-out to convert to FCF later.
Reverse DCF / implied growth. At $402 and ~21x forward, the market implies roughly low-teens long-run earnings growth — below TSMC's delivered +36% revenue / +77% EPS and the raised >40% FY26 guide. The price embeds less growth than the fundamentals are currently producing; the gap is the AI-capex-sustainability discount the market applied post-print.

Embedded Optionality / Free Upside

Framing: the in-production foundry justifies most of the $402; the N2/packaging/photonics options are upside largely for free. Tilt: +4 to Valuation.

Analyst price-target consensus. Consensus $586 (+45.7%) · median $600 (+49.1%) · high $700 (+74%) · low $480 (+19.3% — even the low is above price). Targets were raised sharply post-earnings (last-month avg $590 vs last-year avg $471). Price >20% below consensus → strong valuation support.
Analyst grades & FMP cross-reference. 18 Buy / 7 Hold / 0 Sell (n=25) → 72% bullish, consensus Buy; all recent firm actions (Barclays OW, DA Davidson Buy, Susquehanna Positive, 16-17 Jul) were "maintain" post-print — stable, no downgrades. FMP health rating "B+" (ROE 5/5, ROA 5/5; dragged only by P/E 2 and P/B 1 — the valuation richness, not a quality flaw).
5

Pillar Detail: Underlying Drivers

The dominant external force the stock is tethered to, scored 0–100. A context pillar: it does not change the base signal — it feeds amplification (tailwind ≥65 can lift BUY→STRONG BUY; headwind ≤35 can push SELL→STRONG SELL).
Primary Driver
AI / HPC advanced-node foundry demand cycle
88
Strong Tailwind — amplification-eligible (lifts long BUY → STRONG BUY)

Primary driver: the AI / HPC advanced-node demand cycle (secondary: smartphone/auto recovery; and the Taiwan geopolitical tail as an offsetting risk). TSMC is the leveraged, upstream bet on AI compute — every leading-edge AI accelerator is fabbed here. This is not a commodity-price driver, so no spot-trend overlay applies; it is scored on demand signals, utilisation and capex guidance. The driver stayed a Strong Tailwind (88) — the Q2 call raised the outlook and cited robust demand through 2029-2030.

HorizonReadScore
Historical (12-24m)AI/HPC went from emerging to the dominant revenue driver; advanced-node utilisation >90% throughout85
Current stateStrong: FY26 guide raised to >40% growth, capex raised to $60-64B, +$100B Arizona, CEO calls demand "extremely robust." The only near-term wrinkle is that the capex raise itself stoked an AI-capex-sustainability debate (BofA survey: 45% of managers see an AI bubble as the top risk)92
Forward (6-12m)Demand visibility strong (guided through 2029-30); the risk is a hyperscaler capex guide-down (watch the 22-30 Jul mega-cap prints) puncturing the AI-capex narrative82

Driver score: 88 / 100 — Strong Tailwind. Weighted (0.25/0.50/0.25): 85·0.25 + 92·0.50 + 82·0.25 = ~88.

Amplification eligibility: at 88 the driver is a Strong Tailwind (≥65) — eligible to amplify. Combined with the long-horizon macro Tailwind (§6), it lifts the base Long BUY → STRONG BUY. It does not amplify Short (HOLD never amplifies) or Medium (medium macro pressure is Neutral, not Tailwind). Thesis-invalidation floor: a sustained cut in hyperscaler AI-capex guidance and advanced-node utilisation falling below ~80% would break the tailwind — the dial to watch is the mega-cap capex commentary, not TSMC's own (already-raised) numbers.
6

Pillar Detail: Economic Alignment

How the current economic climate sits relative to this stock, read from the latest Macro-Economic report. Classifies the macro pressure (Tailwind / Neutral / Headwind) — the second amplification input — and frames a long entry as Trend-Following or Contrarian with a 0–100 conviction.
Stance · Pressure
Contrarian (short) / Trend-Following (long) · Headwind (S) → Neutral (M) → Tailwind (L)
55
conviction

TSMC is not on the 2026-07-20 macro Economic Watchlist (energy/defensive-tilted), so Economic Alignment is read from the Driver-Sector matrix. Under the Stagflation-lite, energy-supply-shock regime: XLK = Short Underperform / Medium Neutral / Long Outperform, and EM Equities = Short Strong-Underperform / Medium Underperform / Long Outperform (a Taiwan ADR carries both). So the macro pressure is a Headwind short (risk-off, firm USD, Iran/Hormuz EM stress, mega-cap rate drag), Neutral medium, and a Tailwind long (the structural AI/Tech and EM re-rating). Stance: Contrarian near-term (fading a risk-off tape on a washed-out, high-quality name) / Trend-Following long; conviction ~55. Amplification: only the long pressure is a Tailwind, so it enables the Long STRONG BUY; the Neutral medium and Headwind short pressures leave Medium (BUY) and Short (HOLD) un-amplified. The AI-concentration tail is ARMED but NOT triggering (breadth broadening — equal-weight RSP beating SPY), and TSMC does not belong to that US-index / non-operating-earnings cohort, so it does not inherit the concentration bear leg (DNB 2(b) not fired).

Source: watchlist-signal absent → sector-map (GICS Tech → XLK; plus EM-Equities read for a Taiwan ADR) · Macro report 2026-07-20

7

Pillar Detail: Entry/Exit Timing

The risk-reward framework, relative strength vs SPY and the sector ETF, the macro overlay, news-derived sentiment, and the catalyst cluster.
Entry/Exit Timing — Pillar Score
Great business, cheaper price — but the post-earnings tape rolled over
49
confidence 55%
Sub-signalReadScore
MTF trend scoreMonthly & weekly uptrend (the primary trend is intact) / daily weakening + support-breakdown (RSI 39, below SMA20 $437 & SMA50 $425) / hourly strong-down / 15-min down → weighted ~54; tool confluence bearish54
Risk-reward (daily)Price $402 bounced off the $386 post-earnings low; nearest support $384-386, stop below $380 (~1x the $19 daily ATR) — a fairly tight stop, but the name is below its 50-DMA in a fresh downtrend48
Relative strengthStill +80-90% over 52 weeks (a leader), but has rolled over ~-6% over the last month and gave back the June spike to $479. 52-wk position ~52% after the drop52
Macro overlay (Semis, 15% wt)XLK short Underperform under stagflation-lite risk-off; firm USD + Iran/Hormuz EM stress weigh on a Taiwan ADR42
Sentiment (grades + news)All recent firm actions "maintain" (0 downgrades), targets raised to $600 median — constructive; but the stock fell on record earnings (capex/AI-bubble fears), so near-term news tone is mixed-negative55
Catalyst layerOwn earnings just passed (next ~mid-Oct); near-term path risk is the 22-30 Jul AI mega-cap prints + FOMC 29 Jul spilling into the AI-capex debate55

Timing score: 49 / 100 — Neutral, deteriorated from 55. Composition: MTF 0.30 + risk-reward 0.20 + macro 0.15 + sentiment 0.18 + catalyst 0.17. The honest read: this is a classic higher-timeframe-uptrend / lower-timeframe-pullback — a potential buy-the-dip setup for Medium/Long, but the daily tape is a confirmed downtrend with a support breakdown, so the near-term timing edge is not there yet. This is what caps the Short at HOLD (see banner).

Position-risk: nearest logical stop is below the $384 post-earnings shelf (structural: the $352 200-DMA), ~1x the daily ATR ($19) from price for the tight stop. A reclaim of the ~$425 50-DMA on volume is the technical confirmation the Short is waiting for; a loss of $384 opens the $352 200-DMA test.
8

Economic Event Risk

High-impact macro releases in the next 14 days that could swing this stock, plus the last 7 days of surprises.

Upcoming events (next 30 days)

DateEventImpactForecastPreviousRelevant?Why
2026-07-22Alphabet (GOOGL) Q2 earningsHighCapex guided up (~$185B)✅ YesAI-capex read-through — a hyperscaler capex cut/ROI caution is the direct falsifier of TSMC's demand tailwind
22-30 JulMega-cap AI prints (Meta/MSFT/AMZN)High✅ YesThe AI-capex-sustainability question that sank TSM post-print resolves here
2026-07-29FOMC Rate Decision (Warsh)HighHold 3.50-3.75%3.75%⚠ MediumHigher-for-longer + firm USD is a headwind for EM/Taiwan tech multiples
2026-07-30US Q2 GDP (advance)Medium~2.0% ann.⚠ MediumSets the risk-on/off tape; growth scare would pressure high-beta semis

Recent surprises (last 7 days)

DateEventActualForecastSurpriseImpact
2026-07-15/16TSMC Q2 2026 earningsBeatGM 67.7%, EPS +77%, FY guide >40%beat + raiseRecord quarter — but stock FELL ~4.6% on the $60-64B capex raise (AI-spend-sustainability fear)
2026-07-17Michigan Consumer Sentiment (Jul)54.451.0aboveFirmer US consumer — mild risk-on, partly offsets the tech sell-off
2026-07-18/20Iran/Hormuz escalationStrait disruptedrisk-offUS strikes on Iran nuclear sites; Brent ~$88 — firm USD + EM risk-off weigh on a Taiwan ADR

The stock-specific catalyst (Q2 earnings) has passed — a record quarter that the market sold on capex/AI-bubble fears, not on the numbers. The live near-term risk is now read-through: the 22-30 Jul AI mega-cap prints will either validate or puncture the AI-capex narrative that TSMC is levered to, and the FOMC (29 Jul) + firm-USD/Iran risk-off set an unfavourable EM tape. Semis are medium macro-sensitivity, so no WAIT-for-event override fires — but the next update is scheduled the trading day after the mega-cap cluster (31 Jul).

9

Multi-Timeframe Technical Analysis

Trend, RSI and breakout status across monthly / weekly / daily / hourly / 15-minute, with a confluence verdict.
TimeframeTrendDirectionRSIMACDKey S/RBreakoutVol
MonthlyUptrend ↑Bullish66+, hist +S: 134 / R: 226Resist breakout0.7x
WeeklyUptrend ↑Bullish55+, hist turningS: 224-275 / R: 479Resist breakout0.2x
DailyWeakening →Bearish39−, hist −5.8S: 384-386 / R: 442-479Support breakdown1.3x
HourlyStrong down ↓Bearish50−/flatS: 386-397 / R: 408-428Support breakdown
15-minDowntrend ↓Bearish52−/flatS: 400 / R: 405-410Support breakdown
Confluence: Bearish near-term within a bullish primary trend · MTF Score 54

A textbook higher-timeframe-uptrend / lower-timeframe-pullback. The monthly and weekly trends remain uptrends (the primary trend is intact and both are above their rising 50-week/50-month averages), but the daily has broken down post-earnings — price $402 sits below the SMA20 ($437) and SMA50 ($425), daily RSI 39, MACD histogram −5.8, and the tool flags a support breakdown across daily/hourly/15-min (confluence bearish). Key levels: the $384-386 post-earnings shelf is first support (loss opens the $352 200-DMA); a reclaim of the ~$425 50-DMA on volume would confirm the primary uptrend has resumed. This bearish lower-timeframe confluence is the basis for the Short technical-confirmation cap — the dip is not yet a confirmed entry.

10

Price Chart (6-Month Daily)

A 6-month daily close line with SMA50 and key support/resistance — the visual companion to the MTF table.

6-month daily close (orange = SMA50). The June spike to $479 gave way to a post-earnings breakdown: price $402 is below the $425 SMA50 with a support breakdown, but well above the $352 200-DMA. $384-386 is the post-earnings shelf; $600 is the raised analyst median target.

11

Scenario Summary

Bull / Base / Bear 12-month price paths with triggers and probability weights.

Bull $560 (30%, 12m)

The AI-capex fear proves overblown: hyperscalers keep guiding capex up, N2 ramps on schedule, gross margin holds >65%, and the capex build converts toward FCF. The multiple re-rates back toward 26-28x forward and the stock closes the gap to the raised $600 median target (+39%).

Base $480 (50%, 12m)

FY26 delivers the guided >40% growth, ~23x forward holds, and the stock converges part-way to the Street's $586 consensus as the capex-sustainability worry fades but does not fully resolve. The probability-weighted centre of gravity (+19%).

Bear $340 (20%, 12m)

AI-capex digestion + geopolitics: a hyperscaler cuts capex guidance and the AI-spend narrative deflates, the $60-64B capex dilutes margins, Intel 18A / Samsung take some advanced-node share, and a firm-USD / Iran risk-off tape (or a Taiwan/China escalation) compresses the multiple toward and through the $352 200-DMA (−15%).

Probability-weighted 12-month value ≈ 0.30·$560 + 0.50·$480 + 0.20·$340 = ~$476 (+18%). Favourably skewed: bear downside ~−15% vs base +19% / bull +39%, with even the low analyst target ($480) above the current price — positive expected value, but with a genuine AI-capex-digestion tail and the standing Taiwan geopolitical risk.

12

Entry / Exit Rules

Three independent entry paths (Fundamental · Technical · Catalyst) and three exit triggers (Stop-Loss · Thesis · Profit-Target). Any one entry path is a valid entry — the more that agree, the larger the position the conviction ladder suggests. Exits are graded by severity, not count.

How to read this — the Conviction Ladder

The three entry groups are alternative paths to a buy, not a checklist. A group counts only when all its sub-conditions hold. How many groups are satisfied sets the suggested size — it does not gate whether you may enter: 1 group = Half-Size (a valid starter/scale-in), 2 = Full-Size, 3 = Over-Size (highest conviction); 0 = Wait (no path open yet). A strong overall signal can still read Wait here when the stock is well above its entry zones — that flags "good business, no entry edge right now," not a contradiction. Exits are graded by severity of what is live, not by a count: a hard stop is an Exit on its own.
Entry conviction: Half-Size1 of 3 groups met — one path open — starter / scale-in

Fundamental — MET

Cheap on forward earnings + reverse-DCF, driver a Strong Tailwind, no earnings blackout.
✅ Price $402 < fair-value estimate ~$480
✅ No earnings within 7 days (Q2 reported 15-16 Jul; next ~mid-Oct)
✅ Underlying-Driver score ≥ 50 (88)

Technical — not MET

UNMET — post-earnings breakdown: daily below the 50-DMA, MACD histogram negative, no volume-backed reclaim.
⛔ Daily close above SMA50 ($425) on >1.5x volume (breakout) OR a tested higher-low bounce off $384-386 support — neither confirmed (price $402 < $425; daily is a downtrend/support-breakdown)
✅ RSI 35-65 (daily 39)
⛔ MACD daily histogram positive ≥2 days (currently −5.8)

Catalyst — not MET

Earnings delivered, but the reaction was NEGATIVE — the confirming condition failed.
⛔ Post-earnings move >+5% with guidance raised on >2x volume (guidance WAS raised, but the stock fell ~4.6% — condition not met)

Forecast: Only the Fundamental path is met → 1 of 3 = Half-Size (a starter / scale-in). The Short is capped at HOLD because neither Technical nor Catalyst is met (short technical-confirmation cap). The Technical path opens on a daily reclaim of the ~$425 50-DMA on >1.5x volume, or a tested higher-low off $384-386forecast: ~2-4 weeks, Moderate confidence, catalyst-dependent on the 22-30 Jul AI mega-cap prints stabilising the AI-capex tape. If those disappoint, a test of $384 (then the $352 200-DMA) is the more likely near-term path.

Exit action: Holdno exit trigger is live — hold the position

Stop-Loss — not LIVE

⛔ Two daily closes below $380 (below the $384-386 post-earnings shelf; structural $352 200-DMA)

Thesis Invalidation — not LIVE

⛔ Sustained hyperscaler AI-capex guide-down AND advanced-node utilisation falling below ~80% (the driver floor)
⛔ Competitive break: Intel 18A / Samsung establish a credible second leading-edge foundry source at scale and take material advanced-node share

Profit-Target — not LIVE

⛔ Price into the $600 median target with RSI > 70 and no quality improvement to justify it

Forecast: Stop ($380) is ~5.5% below price; a break needs the AI-capex fear to harden into a hyperscaler guide-down or a Taiwan escalation — possible on a bad 22-30 Jul mega-cap print, else unlikely in 4-6 weeks. Profit-target (≥$600 + overbought) is >49% away — not near-term. No exit trigger is live; action = Hold.

Imagine you act at the current price of $402.30 · as of 20 Jul 2026

What if you bought now?

You are risking ~6% (to the ~$380 stop) / ~15% (bear) to gain ~19% (base) and ~39% (bull) on a wide-moat AI toll-booth that just de-rated on a record quarter.

What you're risking: the drawdown to the ~$380 stop (−5.5%) and, in the bear case, ~−15% to $340 if the AI-capex narrative deflates or Taiwan risk flares; plus you'd be buying below the 50-DMA in a confirmed daily downtrend (the Technical path is NOT met) into the 22-30 Jul mega-cap / FOMC event window. What you're gaining: a Quality-88 near-monopoly foundry at ~21x forward and 45-49% below the Street (even the low target is +19%), a record 67.7% gross margin and a raised >40% growth guide, plus the free N2/advanced-packaging optionality — you start capturing the base +19% / bull +39% immediately. Read: for Medium/Long this is a fair scale-in zone (hence BUY / STRONG BUY); for the Short it pays to wait for the $425 reclaim or a stabilised AI-capex tape — acting now buys event risk with the daily trend against you.

What if you sold now?

You'd be giving up ~19-49% of upside to sidestep an ~15% bear.

What you're giving up: +19% to the $480 base, +49% to the $600 median, +74% to the $700 high, plus the AI-secular compounding and the N2/packaging optionality — selling well below every analyst target and below fair value on a name whose own numbers just beat and raised. What you're protecting: the ~15% bear drawdown to $340 if AI-capex digestion or a Taiwan shock hits, and the near-term downtrend/event path-risk. Is any exit rule live? No — price is above the $380 stop, far from the $600 profit-target, and no thesis-invalidation condition is met. Read: no mechanical reason to sell; for a holder this is a hold/accumulate zone — the deterioration is in the tape (Short), not the business.

13

Position Sizing Context

Illustrative portfolio math (not advice) translating conviction into an allocation given risk-per-share and volatility.

The §12 Conviction Ladder reads Half-Size (1 of 3 entry paths met) — only the Fundamental path is open; the Technical and Catalyst paths need the tape to turn. No user allocation or portfolio role was provided, so a specific portfolio % is not computed. Specify your allocation and role for sizing guidance.

Volatility context: daily ATR ~$19 = ~4.8% of price. Beta ~1.37 (well above market risk). 52-wk range $224-$479; the stock is ~16% off its June high. A staggered entry suits the unconfirmed tape: a starter tranche now, add on a $425 50-DMA reclaim or a stabilised AI-capex read, and a third near the $384 shelf / $352 200-DMA if it pulls back. Note the standing Taiwan geopolitical tail argues for sizing this below a comparable non-Taiwan name.

14

Calibration Snapshot

Machine-readable snapshot of every score, level and signal, saved alongside the HTML so the next run can compute deltas.
{
  "ticker": "TSM",
  "exchange": "NYSE",
  "exchange_ticker": "NYSE:TSM",
  "api_ticker": "TSM",
  "storage_ticker": "TSM",
  "isin": "US8740391003",
  "date": "2026-07-20",
  "version": "v6",
  "company": "Taiwan Semiconductor Manufacturing Company Limited",
  "currency": "USD",
  "reported_currency": "TWD",
  "security_note": "NYSE ADR (1 ADR = 5 TWSE ordinary shares); analysed as the USD ADR.",
  "analysis_status": "on-going",
  "finder_ticker": "TSM",
  "finder_exchange": "\ud83c\uddfa\ud83c\uddf8 NYSE",
  "section": "Emerging-Market Equities",
  "mode": "batch",
  "user_horizon": null,
  "user_allocation_pct": null,
  "portfolio_role": null,
  "lifecycle_stage": "growth",
  "sector": "Technology",
  "sub_industry": "Semiconductors \u2014 Advanced-Node Foundry",
  "gics_sector": "Information Technology",
  "country": "Taiwan",
  "price_at_rating": 402.3,
  "price_asof": "2026-07-20",
  "market_cap_usd": 2086520904000,
  "beta": 1.37,
  "fifty_two_week_range": "223.7-479",
  "signal_short": "HOLD",
  "signal_medium": "BUY",
  "signal_long": "STRONG BUY",
  "primary_signal": "STRONG BUY",
  "base_signals_pre_amplification": {
    "short": "HOLD",
    "medium": "BUY",
    "long": "BUY"
  },
  "amplification_applied": {
    "short": "none (HOLD never amplifies)",
    "medium": "none (medium macro pressure Neutral, not Tailwind)",
    "long": "BUY->STRONG BUY (driver 88 Strong Tailwind + long macro Tailwind XLK/EM long O)"
  },
  "short_entry_confirmed": false,
  "short_cap_reason": "Short technical-confirmation cap: base BUY-accumulate fired on the Fundamental group only; Technical & Catalyst entry groups unmet (post-earnings daily downtrend, below 50-DMA, MACD negative, support breakdown; earnings reaction was negative). Buy on confirmation of a ~$425 50-DMA reclaim or a stabilised AI-capex tape.",
  "quality_score": 88,
  "valuation_score": 64,
  "timing_score": 49,
  "driver_score": 88,
  "quality_detail": {
    "gross_margin_latest_q": 0.677,
    "operating_margin_q": 0.6,
    "net_margin_q": 0.556,
    "roe": 0.362,
    "interest_coverage": 176.1,
    "current_ratio": 2.49,
    "debt_to_equity": 0.17,
    "net_cash": true,
    "industry_benchmark_name": "Gross Margin + Capacity Utilisation",
    "industry_benchmark_value": "GM 67.7% + util >90%",
    "industry_benchmark_score": 93,
    "moat_score": 82,
    "roic_percentile_vs_peers": 90,
    "capital_allocation": 86,
    "management_skin_in_game": 60
  },
  "valuation_detail": {
    "pe_trailing": 31.0,
    "pe_forward_ntm": 21.0,
    "peg_forward": 1.05,
    "price_to_book": 10.2,
    "price_to_sales": 14.7,
    "fcf_yield": 1.7,
    "dividend_yield": 0.95,
    "implied_growth_rate": 13.0,
    "consensus_growth_rate": 30.0,
    "historical_valuation_decile": 5,
    "fair_value_estimate": 480
  },
  "timing_detail": {
    "mtf_confluence": 54,
    "risk_reward_score": 48,
    "relative_strength_vs_spy": 80.0,
    "relative_strength_1mo": -6.0,
    "catalyst_clustering_score": 55,
    "dynamic_macro_weight": 0.15,
    "rsi_daily": 39.1,
    "atr_daily": 19.19,
    "sma50_daily": 425.33,
    "sma200_daily": 351.74
  },
  "nonop_pct_of_net_income": 14,
  "clean_pe": 31.0,
  "clean_peg": 1.05,
  "val_band": "attractive",
  "warranted_multiple": 28.0,
  "actual_multiple": 21.0,
  "warranted_ratio": 0.75,
  "val_multiple_basis": "forward P/E (NTM); trailing ~30-31x = Fair-rich at the 28x guardrail line",
  "discount_rate_r": 9.05,
  "risk_free_10y": 4.55,
  "g_near": 0.15,
  "g_term": 0.03,
  "sector_guardrail": 28.0,
  "competitive_share_trajectory": "gaining",
  "competitive_threat_level": "moderate",
  "economic_alignment_stance": "Contrarian (short) / Trend-Following (long)",
  "economic_alignment_conviction": 55,
  "economic_alignment_pressure": "Headwind (S) / Neutral (M) / Tailwind (L)",
  "economic_alignment_pressure_short": "Headwind",
  "economic_alignment_pressure_medium": "Neutral",
  "economic_alignment_pressure_long": "Tailwind",
  "economic_alignment_source": "sector-map",
  "macro_report_date": "2026-07-20",
  "sector_signal_xlk": {
    "s": "U",
    "m": "N",
    "l": "O"
  },
  "em_equity_signal": {
    "s": "SU",
    "m": "U",
    "l": "O"
  },
  "analyst_consensus_target": 586,
  "analyst_target_high": 700,
  "analyst_target_low": 480,
  "analyst_target_median": 600,
  "analyst_target_upside_pct": 49.1,
  "analyst_grades_consensus": "Buy",
  "analyst_bullish_pct": 72,
  "analyst_coverage_count": 25,
  "fmp_rating": "B+",
  "fmp_overall_score": 3,
  "recent_upgrades_30d": 0,
  "recent_downgrades_30d": 0,
  "overall_confidence": 55,
  "fair_value_est": 480,
  "stop_loss": 380,
  "target_price": 480,
  "scenario_base_target": 480,
  "scenario_bull_target": 560,
  "scenario_bear_target": 340,
  "entry_groups_met": 1,
  "entry_conviction": "Half-Size",
  "exit_groups_live": 0,
  "exit_action": "Hold",
  "hard_gate_state": "caution",
  "gates_triggered": [],
  "gates_caution": [
    "Taiwan/China geopolitical binary tail (Gate 5 caution \u2014 structural, not a dated ruling)",
    "Valuation Ceiling caution \u2014 trailing P/E ~30-31x at the 28x Semis guardrail line; forward ~21x is not expensive so ceiling does not fire",
    "Capex-driven FCF yield ~1.7% (rich on cash) and rich P/B/P/S"
  ],
  "do_not_buy_triggers": [],
  "dnb_note": "AI-concentration tail ARMED but not triggering (breadth broadening); TSMC is a foreign foundry with clean operating earnings, NOT in the S&P-500-concentration / non-operating-gains cohort, so it does not inherit the concentration bear leg \u2014 DNB Trigger 2(b) NOT fired. Not in the Anchor's Expensive band on forward earnings.",
  "next_update_date": "2026-07-31",
  "next_check_date": "2026-07-31",
  "next_update_basis": "AI mega-cap prints (22-30 Jul, AI-capex read-through) + FOMC 29 Jul + Q2 GDP 30 Jul cluster +1 trading day (31 Jul); own Q3 print ~mid-Oct",
  "last_updated_human": "Jul 20, 2026",
  "prior_report": "calibration-TSM-20260616-1703.json",
  "prior_primary": "STRONG BUY",
  "changes_note": "Signals unchanged HOLD/BUY/STRONG BUY. Q2 reported (record: rev +36%, GM 67.7%, EPS +77%, guide raised >40%, capex raised $60-64B) but stock fell on capex/AI-bubble fear. Valuation 57->64 (fwd P/E ~21x after de-rate; median target 467->600). Timing 55->49 (post-earnings daily breakdown). Quality/Driver/Econ flat. Entry Half-Size. Gate caution (Taiwan tail; earnings gate now clear). Price -5.5%. DNB 2(b) NOT fired (clean earnings, not in AI-concentration cohort)."
}

Mode-B refresh (2026-07-20). Signals HOLD / BUY / STRONG BUY — unchanged from 2026-06-16. Short stays HOLD via the short technical-confirmation cap (post-earnings tape rolled over: daily downtrend + support breakdown, Technical & Catalyst entry groups both unmet, only Fundamental met). Medium BUY: Quality 88 + Valuation 64 (now Attractive on ~21x forward after the de-rate + 45-49% Street upside) carry the matrix even with Timing Neutral (49); Neutral medium macro pressure blocks STRONG BUY. Long STRONG BUY: driver 88 Strong Tailwind + long-horizon Tech/EM macro Tailwind amplify the base BUY. Q2 2026 reported (was estimated last run): revenue +36%, GM 67.7% record, EPS +77%, FY guide raised >40%, capex raised $60-64B + $100B Arizona — the stock fell on capex/AI-bubble fears despite the beat. Earnings clean (non-op ~14% of net, below thresholds); TSMC not in the S&P-concentration / non-op AI cohort, so DNB 2(b) not fired. Hard gate CAUTION (Taiwan tail + trailing-multiple/capex-FCF), no DNB triggers.

15

Data Sources & Methodology

Audit trail of every data source: fully available (✓), fallback (⚠), or failed (✗), plus provenance-based confidence haircuts.
Data Source Status
get_company_profile / get_stock_snapshot price $402.30, mkt cap $2.09T, beta 1.37, ISIN, ADR
get_financial_ratios GM 61.9% TTM, net margin 47% TTM, ROE 36%, coverage 176x, P/E 31x, FCF yield ~1.7%
get_income_statement 6 quarters TWD; Q2'26 rev +36%, GM 67.7%; earnings-quality decomposition (non-op ~14% of net)
get_multi_timeframe_analysis 5 timeframes; confluence bearish (daily/hourly/15m breakdown within monthly/weekly uptrend)
get_stock_prices 125 daily bars for chart + SMA50
get_analyst_estimates FY26-29 rev/EPS; multi-year ~30% EPS CAGR
get_price_target_consensus / summary consensus $586, median $600, high $700, low $480; raised post-earnings
get_stock_grades / grades_consensus 18 Buy / 7 Hold; all recent 'maintain' (16-17 Jul)
get_ratings_snapshot FMP B+ (ROE/ROA 5/5; P/E 2, P/B 1)
get_polygon_news / search_financial_news Q2 reaction — record quarter, stock fell on $60-64B capex raise + AI-bubble fear; Intel 18A Apple/Tesla wins
get_economic_calendar / get_economic_series (DGS10) 10-Y 4.55% (17 Jul) for the anchor; 22-30 Jul mega-cap + FOMC cluster
MacroDriver-state-20260720 Stagflation-lite; XLK U/N/O, EM SU/U/O; TSM not in watchlist → sector-map; AI tail armed-not-triggering (not inherited)
get_earnings_calendar / get_related_tickers earnings calendar empty (Q2 date confirmed from the filed 15-16 Jul income statement); related-tickers errored — competitors sourced via news/knowledge
Impact on scores: Strong MCP coverage; Q2 numbers are hard (filed income statement). Two minor gaps: get_related_tickers errored (competitor set built from news + domain knowledge — Intel Foundry, Samsung, hyperscaler custom silicon), and peer semis medians / relative-strength are partly estimated (Valuation/Timing confidence trimmed accordingly). Overall confidence 55% = min(Quality 82, Valuation 78, Timing 55) — Timing confidence the binding constraint given the fresh downtrend + risk-off tape.
DISCLAIMER: This is a quantitative framework for educational purposes only. It is not financial advice. Always do your own research and consult a licensed financial advisor before making investment decisions.