valuation / 2330.TW
2026-05-24 · Comps · P/E · high · source: weilab/model-builder@2026-05-24
| cash_m | 2,127,000 |
| debt_m | 940,000 |
| currency | TWD |
| eps_path | [{"eps":{"low":70.1,"mid":78.08,"high":89.75},"year":"CY26E"},{"eps":{"low":82.72,"mid":92.13,"high":105.91},"year":"CY27E"},{"eps":{"low":97.61,"mid":108.71,"high":124.97},"year":"CY28E"}] |
| pe_peers.GFS | 19.52 |
| pe_peers.INTC | 25.77 |
| pe_peers.0981.HK | 21.39 |
| pe_peers.2303.TW | 11.8 |
| pe_peers.000660.KS | 17.65 |
| pe_peers.005930.KS | 19.76 |
| pe_applied | 25.53 |
| pe_terminal | 22 |
| eps_estimate.low | 70.1 |
| eps_estimate.mid | 78.08 |
| eps_estimate.high | 89.75 |
| period_label | CY26E |
| pe_peer_stats.p25 | 18.12 |
| pe_peer_stats.p75 | 20.98 |
| pe_peer_stats.median | 19.64 |
| revenue_cagr_5y | 18.00% |
| hbm_premium_note | Applied P/E of 25.53x = peer median 19.64x × (1+30% premium). Premium reflects: (a) CoWoS-L advanced packaging capacity monopoly through 2027 H2 (Samsung/Intel cannot replicate); (b) N2 process node ~1-year lead vs Samsung Foundry SF2 and Intel 18A; (c) demonstrated pricing power — CoWoS-L 2026 wafer pricing up ~20% YoY per TSMC IR commentary. Market is paying ~47% premium (27.5/19.64-1), which exceeds the defensible 30% — implies further upside requires either applied multiple > 30x (TSM-ADR style 35x territory) or EPS beat above TWD 90. |
| spot_price_source.GFS | stale_estimate |
| spot_price_source.INTC | weilab.stock_quotes@2026-05-22 |
| spot_price_source.0981.HK | stale_estimate |
| spot_price_source.2303.TW | stale_estimate |
| spot_price_source.2330.TW | weilab.stock_quotes@2026-05-22 |
| spot_price_source.000660.KS | weilab.stock_quotes@2026-05-22 |
| spot_price_source.005930.KS | weilab.stock_quotes@2026-05-22 |
| shares_outstanding_m | 25,930.4 |
| pe_terminal_rationale | Compresses from 25.53x applied (which is peer median 19.64x × 1.30 premium) to 22.0x by FY28 — reflecting partial erosion of CoWoS-L monopoly premium as Samsung/Intel ramp competing advanced packaging through 2027-2028 and N2 process lead narrows. NOT full compression to peer median (19.64x) because N2P/N1.4 transition + 3D IC moat persists. Mean-reversion interpolation: FY+1=25.53, FY+2=23.77 (i), FY+3=22.00. |
| ebitda_margin_terminal | 55.00% |
| forward_pe_reconciliation.note_md | Top-down EPS (TWD 78.08) is within -4.78% of market-implied EPS (TWD 82.00 at consensus 27.5x Fwd P/E) — well inside the ±20% tolerance band. Headline anchors to top-down × peer-relative applied P/E. The fact that market is paying ~27.5x while peer-relative-with-premium only justifies ~25.5x is the key finding: TSMC monopoly premium is already priced in. |
| forward_pe_reconciliation.anchor_choice | top_down |
| forward_pe_reconciliation.market_fwd_pe | 27.5 |
| forward_pe_reconciliation.divergence_pct | -4.78 |
| forward_pe_reconciliation.top_down_eps_mid | 78.08 |
| forward_pe_reconciliation.market_implied_eps | 82 |
[audit 2026-05-24] manual SQL patch: added eps_path (CY27E/CY28E rolled at 18% revenue_cagr_5y) + pe_terminal=22.0 (partial compression from 25.53x toward peer median 19.64x). No sub-agent re-run; existing analysis preserved. TSMC peer-relative Fwd P/E CY26E with +30% premium for CoWoS-L capacity monopoly and N2 process lead. Top-down EPS (TWD 78.08) reconciles within -4.8% of market-implied EPS (TWD 82.0 at consensus 27.5x), so headline anchors to top-down per discipline rule. Mid implied TWD 1,993 is ~-11.6% vs TWD 2,255 spot — market is pricing TSMC at ~47% premium to peer median, well above the 30% defensible premium. TSMC is the foundry multiple-setter; market enthusiasm is already in the price.
