short_entry_confirmed still false (below the 50-DMA); the half position is the risk control.Taiwan Semiconductor (TSMC) is the world's largest dedicated contract chip manufacturer — a "pure-play foundry" that builds the physical silicon designed by others (Apple, Nvidia, AMD, Qualcomm, Broadcom) but sells no chips of its own. Its core business is running the most advanced fabrication plants on earth, turning circuit designs into finished wafers at nodes (N3, N2) that essentially no rival can match at volume. What sets it apart is a near-monopoly on leading-edge manufacturing: it holds roughly 90%+ of the advanced-node foundry market, a structural cost and yield advantage from unmatched scale, and customers so deeply designed-in that switching is impractical. For a reader, think of it as the indispensable factory behind the entire AI and smartphone chip supply chain — with the one large caveat that its fabs sit in Taiwan, at the centre of a geopolitical fault line. Analysed here as the USD-denominated NYSE ADR (1 ADR = 5 Taiwan ordinary shares); the company reports in New Taiwan dollars.
Lifecycle: Growth-stage dominant leader. Revenue +36% YoY (Q2 2026 TWD 1,270.4bn), record gross margin 67.7%, operating margin 60.3%, net margin 55.6%, ROE ~40%, ROA ~19%. A rare combination of hyper-scale growth and cash-cow profitability.
| Sub-signal | Reading | Score |
|---|---|---|
| Revenue trajectory | +36% YoY, guide raised to >40% growth for 2026 (AI-led) | 95 |
| Profitability vs peers | GM 67.7% (record), op margin 60% — top of the foundry universe | 95 |
| Cash generation | Op cash-flow margin ~60%; but FCF yield only ~1.8% (record $60-64bn capex) | 62 |
| Balance-sheet health | Net cash, D/E 0.15, interest coverage ~285×, current ratio 2.46 | 92 |
Moat score 82/100. ROIC top-decile vs semis peers; capital-allocation disciplined (record but demand-backed capex, rising dividend, payout ~24%). Management skin-in-the-game modest (widely-held, state-linked shareholder base) — scored 60.
| Competitor | Position | Share trend |
|---|---|---|
| Intel Foundry (18A) | Behind on external volume + yield | Flat / trailing |
| Samsung Foundry (SF2) | Yield gap at leading edge | Losing at N2 |
| Rapidus (2nm pilot) | Pre-production, 2027 target | N/A |
| SMIC (China) | EUV-blocked; DUV workarounds only | Capped at leading edge |
Warranted-Multiple Anchor. Discount rate r = 4.67% (10-Y, 2026-07-30 macro) + 4.5% ERP + 0.0% risk add-on (Quality ≥ 65) = 9.17%. Growth g_near = 15% (secular-growth cap; consensus ~30% haircut to the sector-achievable ceiling — the anti-hype discipline), g_term = 3%. Two-stage warranted P/E ≈ 27.5× (right at the 28× Semis guardrail line).
| Metric | Value | Read |
|---|---|---|
| Forward NTM P/E | ~21× | Attractive vs warranted 27.5× (ratio 0.76) |
| Trailing clean P/E | ~28-30× | At the guardrail line — not cheap |
| PEG (forward) | ~0.9 | Attractive for the growth |
| P/B | ~9.8× | Rich |
| FCF yield | ~1.8% | Low — record capex |
| Dividend yield | ~0.9% | Payout ~24%, rising |
Analyst cross-check: consensus target $596 (median $600, high $700, low $500 — FMP); Yahoo mean $540/median $534. Grades: 18 Buy / 7 Hold / 0 Sell (consensus Buy, ~72% bullish); Yahoo recommendation strong_buy (18 analysts). FMP health rating B+ (overall 3; ROE 5, ROA 5, P/E 2, P/B 1). No upgrades/downgrades in the last 30 days — grades stable.
Primary driver: the AI/HPC advanced-node demand cycle. Q2 guidance raised to >40% revenue growth for 2026; capex raised to $60-64bn; the CFO (CNBC, 19 Jul) confirmed an accelerated Arizona build-out on "robust customer demand." Advanced-node (N3/N2) utilisation >90%. Secondary: smartphone/auto normalisation; the offsetting tail is Taiwan geopolitics.
| Horizon | Driver state | Score |
|---|---|---|
| Historical | Multi-year AI/HPC ramp, share gains | 85 |
| Current | Order book full, capex raised, util >90% | 92 |
| Forward | N2 ramp 2026-27; near-term AI-capex-digestion fear is a sentiment overlay, not yet an order signal | 82 |
Regime: Stagflation-lite (energy shock re-armed + policy-tight into cooling growth; a narrow, contested, tape-unconfirmed lead). Info-Tech (XLK) signals S=N, M=U, L=O; EM Equities S=U, M=SU, L=N. Short/medium pressure is a tech/EM Headwind (contrarian to buy into). Long pressure stays a Tailwind on the XLK long Outperform (the primary sector map for an Info-Tech name) — but the EM-equities long signal was downgraded O→N this cycle, so long conviction is trimmed (55→50) and flagged. The armed 'S&P-500 concentration / AI earnings-quality unwind' systemic tail is carried in the Bear scenario but does NOT apply as a cohort DNB to a clean-earnings foreign foundry.
Source: sector-map (XLK primary; EM secondary) · Macro report 2026-07-30
The primary (monthly/weekly) trend is still up and confluence reads bullish, but the daily is weakening / in a support-breakdown: price $405 sits below the 20-DMA ($413), 50-DMA ($425) and with a negative daily MACD. Constructive that the daily RSI has recovered from ~39 (20 Jul) to ~46 and the stock is holding the $385-$405 shelf while the broader semiconductor tape is in a stated "bear market."
Risk-reward (daily): support $385/$386 then the 200-DMA ~$357; resistance the 50-DMA $425, then $442 / $450 / the $479 high. Reward to base $475 ≈ +17%; risk to the $380 stop ≈ −6%. Relative strength: +100%+ over 52 weeks (also a marker of extension), but ~−6% on a 1-month relative basis as semis de-rate. Catalysts: none dated inside the window (Q3 print ~15 Oct); FOMC 29 Jul passed. Timing classifies Neutral, not Weak — the higher-timeframe uptrend is unbroken.
| Date | Event | Impact | Forecast | Previous | Relevant? | Why |
|---|---|---|---|---|---|---|
| ~15 Oct 2026 | TSMC Q3 2026 earnings | High | EPS est $4.45 | — | Yes | Next own catalyst; AI-capex / advanced-node guidance |
| Aug 2026 | US CPI / jobs (recurring) | Med | — | — | Indirect | Semis are only medium macro-sensitivity — not a scheduling trigger |
| Date | Event | Actual | Forecast | Surprise | Impact |
|---|---|---|---|---|---|
| 29 Jul 2026 | FOMC decision | — | — | — | On-hold, policy-tight bias — stagflation-lite backdrop |
| 30 Jul 2026 | US Q2 GDP | — | — | — | Cooling growth read — feeds the macro caution |
| 15 Jul 2026 | TSMC Q2 2026 earnings | Beat | — | + | Record rev/margins, guide + capex raised (already priced in the last report) |
No dated impactful event inside the 14-day window; the next own catalyst is the Q3 print ~15 Oct. Recurring macro is not a scheduling trigger for a medium-sensitivity semis name.
| Timeframe | Trend | Direction | RSI | MACD | Key S/R | Breakout | Vol |
|---|---|---|---|---|---|---|---|
| Monthly | Uptrend | Up | 67.0 | + (hist +14.0) | S ~358 / R breakout | Resistance breakout | 1.16× |
| Weekly | Uptrend | Up | 56.1 | − (hist −5.2, cooling) | S 224 / R 479 | Resistance breakout | 1.26× |
| Daily | Weakening | Down | 45.9 | − (−9.85) | S 385 / R 450 | Support breakdown | 1.13× |
| Hourly | Recovering | Up | 51.9 | − (hist −1.7) | S 381 / R 425 | Resistance breakout | 0.52× |
| 15-min | Weakening | Down | 41.7 | ± (flat) | S 399 / R 409 | Support breakdown | 0.37× |
| Confluence: Bullish (higher timeframes) with a weak near-term daily · MTF Score 54 | |||||||
Monthly and weekly remain in confirmed uptrends (confluence bullish), but the daily is below all near-term MAs with a negative MACD and a support-breakdown tag — the classic post-earnings consolidation in a semiconductor pullback. The daily RSI recovering off ~39 and price holding the $385 shelf are the constructive tells; a daily close back above the $425 50-DMA on volume would confirm the Technical entry path.
Indicative recent daily closes (post-Q2 pullback from ~$450 to the $385 shelf, now stabilising ~$405, below the $425 50-DMA). Levels are the load-bearing reference.
N2 ramp ahead of plan, GM holds >67%, AI capex re-accelerates and the semiconductor de-rate reverses; forward multiple re-rates back toward ~26-28×. Roughly the level below the $596 consensus target. Trigger: a Q3 guide-up + breadth broadening (RSP catching SPY).
Continued AI-led growth at a ~22-24× forward multiple; the stock grinds toward the low end of the analyst target range as earnings compound. ~+17% from $405. Trigger: order book stays full, util >90%, no macro shock.
AI-capex digestion / a cyclical foundry pause + Taiwan escalation risk + a strong USD. Critically, this path inherits the armed macro AI-concentration / earnings-quality unwind systemic tail: even though TSMC's own earnings are clean, an index-level AI de-rating drags the whole semis cohort (already in a stated bear market), compressing the forward multiple toward ~14-15× and the price through the ~$357 200-DMA toward $325 (~−20%). Falsification: breadth broadening (RSP catching SPY) and hyperscaler capex guidance holding.
Forecast: Fundamental group already met (the basis for the half-size starter). Technical group ~2-4 weeks IF price reclaims the $425 50-DMA on volume — CONFIDENCE Moderate: higher-timeframe trend is up but the daily MACD is negative and the 50-DMA is falling, so a pullback resets the clock. Catalyst group is date-locked to the Q3 print (~15 Oct) — not before. Net: start half now on the Fundamental path; the second tranche waits on the Technical reclaim or a $385 bounce.
Forecast: Stop unlikely in the next 4-6 weeks — $380 is ~6% below and just under the $385 shelf that has held; it would take an earnings-less breakdown or a macro/geopolitical shock. Thesis-invalidation dials are all quiet (util >90%, capex raised). Profit-trim is the more likely first exit if the AI tape re-accelerates.
Acting now buys a structurally dominant, net-cash, 67%-margin AI compounder on a ~21× forward multiple and a ~0.9 PEG — but into a daily downtrend below the 50-DMA and an armed systemic AI-unwind tail. That trade-off is exactly why the call is a half-size starter, not a full position: you get participation on a name we back on both longer horizons, and keep the second tranche for a $425 reclaim or a $385 bounce.
The audit lesson behind the half-size rule: high-quality, non-expensive names held back only by Neutral timing tended to rise anyway and often re-rated to BUY at a higher price. The half position is the risk control that waiting used to provide.
{
"ticker": "TSM",
"exchange": "NYSE",
"exchange_ticker": "NYSE:TSM",
"api_ticker": "TSM",
"storage_ticker": "TSM",
"isin": "US8740391003",
"date": "2026-07-31",
"version": "v6",
"brand": "TSMC",
"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": 405.27,
"price_asof": "2026-07-31",
"market_cap_usd": 2102492790784,
"beta": 1.25,
"fifty_two_week_range": "223.7-479",
"signal_short": "BUY",
"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 on the base HOLD; quality-starter override converts the base HOLD to a half-size BUY (not an amplification)",
"medium": "none (medium macro pressure Headwind, not Tailwind)",
"long": "BUY->STRONG BUY (driver 88 Strong Tailwind + long macro Tailwind on XLK long O; EM long downgraded O->N is a conviction caveat)"
},
"short_entry_confirmed": false,
"short_hold_reason": "neutral_timing_starter",
"short_cap_reason": "Base HOLD is the High-Quality / Fair-Valuation / Neutral-timing row (L1229). Quality-starter override fires (Quality 88>=65, Valuation 64>=40 not Expensive, medium BUY + long STRONG BUY): issue a half-size starter BUY, scale in on a $425 50-DMA reclaim or a $385 support bounce. Technical & Catalyst entry groups remain unmet (below 50-DMA, negative daily MACD, no dated event).",
"quality_score": 88,
"valuation_score": 64,
"timing_score": 50,
"driver_score": 88,
"quality_detail": {
"gross_margin_latest_q": 0.677,
"operating_margin_q": 0.603,
"net_margin_q": 0.556,
"roe": 0.4,
"interest_coverage": 285.4,
"current_ratio": 2.46,
"debt_to_equity": 0.15,
"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": 30.0,
"pe_forward_ntm": 21.0,
"peg_forward": 0.9,
"price_to_book": 9.8,
"price_to_sales": 14.2,
"fcf_yield": 1.8,
"dividend_yield": 0.9,
"implied_growth_rate": 12.0,
"consensus_growth_rate": 30.0,
"historical_valuation_decile": 5,
"fair_value_estimate": 475
},
"timing_detail": {
"mtf_confluence": 54,
"risk_reward_score": 50,
"relative_strength_vs_spy": 100.0,
"relative_strength_1mo": -6.0,
"catalyst_clustering_score": 52,
"dynamic_macro_weight": 0.15,
"rsi_daily": 45.9,
"atr_daily": 18.77,
"sma50_daily": 425.1,
"sma200_daily": 356.78
},
"nonop_pct_of_net_income": 11,
"clean_pe": 30.0,
"clean_peg": 0.9,
"val_band": "attractive",
"warranted_multiple": 27.5,
"actual_multiple": 21.0,
"warranted_ratio": 0.76,
"val_multiple_basis": "forward NTM P/E (~21x, a 5/12 FY26 + 7/12 FY27 EPS blend) vs warranted 27.5x = ratio 0.76 (Attractive band); trailing clean P/E ~28-30x sits at the 28x Semis guardrail line = Fair-rich, which keeps the pillar SCORE at the top of Fair (64)",
"discount_rate_r": 9.17,
"risk_free_10y": 4.67,
"g_near": 0.15,
"g_term": 0.03,
"sector_guardrail": 28.0,
"competitive_share_trajectory": "gaining",
"competitive_threat_level": "moderate",
"economic_alignment_stance": "Contrarian (S/M) / Trend-Following (L)",
"economic_alignment_conviction": 50,
"economic_alignment_pressure": "Headwind (S) / Headwind (M) / Tailwind (L)",
"economic_alignment_pressure_short": "Headwind",
"economic_alignment_pressure_medium": "Headwind",
"economic_alignment_pressure_long": "Tailwind",
"economic_alignment_source": "sector-map",
"macro_report_date": "2026-07-30",
"macro_regime": "Stagflation-lite",
"sector_signal_xlk": {
"s": "N",
"m": "U",
"l": "O"
},
"em_equity_signal": {
"s": "U",
"m": "SU",
"l": "N"
},
"analyst_consensus_target": 596,
"analyst_target_high": 700,
"analyst_target_low": 500,
"analyst_target_median": 600,
"analyst_target_upside_pct": 47.1,
"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": 58,
"fair_value_est": 475,
"stop_loss": 380,
"target_price": 475,
"scenario_base_target": 475,
"scenario_bull_target": 560,
"scenario_bear_target": 325,
"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 clean P/E ~28-30x at the 28x Semis guardrail line at 52-wk highs; forward ~21x and ratio 0.76 are not Expensive so the ceiling does not fire",
"Capex-driven low FCF yield ~1.8% and rich P/B ~10x"
],
"do_not_buy_triggers": [],
"dnb_note": "AI-concentration / earnings-quality-unwind systemic tail is ARMED (breadth-narrowing) in the 2026-07-30 macro report and is carried in the Bear scenario. But TSMC is a foreign foundry on CLEAN operating earnings (~89% operating; non-op ~11% of pre-tax, a positive contributor this quarter), NOT in the S&P-500-concentration / non-operating-gains cohort, and NOT in the Anchor's Expensive band (forward ~21x, ratio 0.76). DNB Trigger 2(b) requires Expensive AND a materially-applicable cohort tail \u2014 neither holds, so it does NOT fire.",
"next_update_date": "2026-08-14",
"next_check_date": "2026-08-14",
"next_update_basis": "default +14d (no dated impactful event inside window; semis medium-sensitivity so recurring macro is not a trigger); own Q3 print ~15 Oct 2026",
"last_updated_human": "Jul 31, 2026",
"prior_report": "calibration-TSM-20260720-1930.json",
"prior_primary": "STRONG BUY",
"changes_note": "Short HOLD -> BUY (half-size starter) on the first refresh under the 2026-07-26 quality-starter rule (High-Quality, non-Expensive, Neutral timing, medium BUY + long STRONG BUY). Medium BUY and Long STRONG BUY unchanged; long conviction trimmed 55->50 on the macro EM-long downgrade O->N + worse medium tech/EM pressure. Scores: Quality 88, Valuation 64, Driver 88 flat; Timing 49->50. Forward NTM P/E ~21x flat (price flat, no new earnings); higher 10Y (4.55->4.67) trims warranted 28->27.5x so ratio ~0.75->0.76; trailing richness keeps Valuation at the top of Fair. Bear target trimmed 340->325 (deeper cohort de-rating leg). Price 402.30 -> 405.27 (+0.7%). Gate caution held; DNB 2(b) NOT fired (clean earnings, not Expensive)."
}
First TSM refresh under the short-horizon quality-starter rule (adopted 2026-07-26): a High-Quality (88), non-Expensive (Fair 64), Neutral-timing (50) name we already rate BUY (medium) and STRONG BUY (long) no longer sits the Short out flat — the Short becomes a half-size starter BUY. Scores otherwise steady; the macro EM-long downgrade (O→N) trims long conviction but XLK-long Outperform keeps the long STRONG BUY.