Concerns & Risks -- 7/10
A favorable risk/reward on the fundamentals, held out of the top tier by geopolitics. Two of the
strongest favorable criteria are cleared cleanly: China exposure is below 10% (~9% and shrinking) and
the valuation sits below the relevant peer average (forward P/E ~24-26x vs ~30x+ US-semi peers) for
what is unambiguously the highest-quality, highest-share leader in the industry. The catalyst path is
rich and near-term. What keeps it out of the 8-10 zone is the binary ITC §337 import-ban ruling on
~75%-of-revenue North America and the permanent Taiwan cross-strait tail.
Weight: 15%
Valuation
Below Peers
Fwd P/E ~24-26x vs ~30x+
Discount to lower-quality peers
China Exposure
~9%
Down from ~22%
Below 10% bar, de-risking
ITC §337 Case
Binary
Ruling late Jun / Oct 2026
Import-ban tail on ~75% of rev
Geopolitics
Taiwan
Cross-strait concentration
Unhedgeable tail
Valuation -- Primary Metric: Forward P/E
| Metric | FY2026E Basis | Multiple | Peer Avg |
|---|---|---|---|
| Forward P/E (primary) | EPS consensus NT$498.73 (+54% YoY) | ~24-26x | ~30-32x |
| EV/EBITDA (secondary) | Scaling with +37% rev growth | 20.3x TTM → high-teens fwd | ~17-22x |
| P/E (TTM) | Trailing | 32.1x | US semi avg ~72.6x |
| P/S (TTM) | Trailing | 15.1x | n/m (capital-intensive) |
The dominant leader trades at a discount to lower-quality peers.
TSM's forward P/E (~24-26x) sits below the broad US-semiconductor peer average (~30-32x, far below
the ~72x industry-wide screen), and EV/EBITDA is roughly in line to slightly cheap vs foundry peers.
PEG is attractive: ~25x forward P/E against ~54% FY26 EPS growth. For the highest-share, highest-quality
name in the group to trade cheap is the favorable setup the rubric rewards.
China Exposure
| Geography | % of Revenue | Trend |
|---|---|---|
| North America | ~75-76% | Rising (from ~56% in 2020) |
| China | ~9% | Falling sharply (from ~22%), export-control driven |
| Rest of World | ~15-16% | Stable |
China is ~9% of revenue -- below the 10% threshold and shrinking.
Management confirmed export-control restrictions on China AI-GPU demand but stated overall AI growth
remains "very dramatic and very positive" even without China. A capped, shrinking, and largely
de-risked piece of the story.
Key catalysts
| # | Catalyst | Timing / Detail |
|---|---|---|
| 1 | N2 (2nm) Ramp + N2P | HVM started 4Q25; N2P volume 2H26. New higher-priced node with structurally better profitability than N3. Near-term 2-3% GM dilution. |
| 2 | Wafer Price Hikes (N7-N2, ~5-10%) | 2026. Pure pricing power into sold-out capacity; advanced nodes ~74% of wafer revenue. |
| 3 | CoWoS / Advanced Packaging | Capacity ~doubling in 2026. Unlocks more AI wafer demand; ~10% of revenue and rising. |
| 4 | Arizona Giga-Fab Cluster | Ongoing (fab 3 building; fabs 4 + packaging permitting). De-risks US tariff/policy exposure at cost of 2-4% overseas GM dilution. |
| 5 | Q2 2026 Revenue Delivery | July 2026. $39.0-$40.2B guide; another likely top-end clear. |
Regulatory / Political risk
| # | Risk | Severity | Detail |
|---|---|---|---|
| 1 | ITC §337 Import-Ban Case | HIGH | Investigation 337-TA-1443 (ex-UMC 7nm-and-smaller patents). ALJ initial determination ~late June 2026, Commission final ~October 2026. Tail risk of a US import ban on advanced-node chips (Apple/Qualcomm/Broadcom named). Single most acute near-term overhang. |
| 2 | Taiwan Cross-Strait | HIGH | Structural, unquantifiable, ever-present geographic concentration -- the foundational bear case for the entire equity. |
| 3 | Tariffs / Export Controls | MEDIUM | Management flags tariff uncertainty in price-sensitive segments but reports no observed change in customer behavior. Overseas fabs (AZ/Japan/Germany) are the policy hedge, at 2-4% GM dilution. |
| 4 | AI Concentration / Overcapacity | MEDIUM | HPC/AI is ~61% of mix; the whole thesis rides one capacity cycle that could pause. $52-56B 2026 capex builds for 2028-2029 demand. |
| 5 | Antitrust | LOW | Dominant share invites scrutiny, but no active foundry antitrust action; pure-play model and customer dependence complicate any case. |
Bull case
| # | Factor | Detail |
|---|---|---|
| 1 | Dominant Leader, Cheap | ~70% foundry leader at forward P/E ~24-26x, below lower-quality peers, with ~54% FY26 EPS growth. |
| 2 | Record and Rising Margins | Q1'26 gross margin 66.2%, operating margin 58.1%; pricing power into sold-out advanced-node capacity. |
| 3 | Multi-Year Catalyst Stack | N2 + CoWoS + AI accelerator ramp; management raised the AI CAGR to mid-to-high 50s%. |
| 4 | China De-Risked | China down to ~9% and falling; growth increasingly North America / AI-driven. |
| 5 | US-Fab Policy Hedge | Arizona giga-fab cluster credibly hedges tariff/policy exposure while the market still discounts geopolitical risk. |
Bear case
| # | Factor | Detail |
|---|---|---|
| 1 | ITC Import-Ban Binary | The late-June / October 2026 ruling is a binary near-term overhang on ~75% of revenue. |
| 2 | Taiwan Concentration | Cross-strait risk is unhedgeable and existential to the equity. |
| 3 | AI Cycle Concentration | HPC/AI ~61% of mix concentrates the thesis on one capacity cycle that could pause or reverse. |
| 4 | Margin Bar Raised | N2 + overseas-fab dilution (~4-6% GM headwind in 2026) plus high-teens depreciation growth after a peak-quality 66% print. |
| 5 | Spending Ahead of Revenue | $52-56B 2026 capex builds for 2028-2029 demand the Street fears could become overcapacity. |
Score rationale
Score of 7/10 reflects a favorable risk/reward on the fundamentals, held out of the top tier by acute geopolitical and regulatory overhang. The rubric's two strongest favorable criteria are cleared cleanly, and the offsets are external tail risks rather than business-quality problems.
What supports the 7 (favorable): China exposure below 10% (~9% and shrinking) (+). Valuation below the relevant peer average -- forward P/E ~24-26x vs ~30x+ US-semi peers -- for the highest-quality, highest-share leader (+). Rich, near-term catalyst path (N2/N2P ramp, ~5-10% wafer hikes, CoWoS doubling, quarterly beats) (+). No real regulatory reshoring overhang; the US-fab build is a policy hedge (+).
What keeps it out of the 8-10 zone: The binary ITC §337 import-ban ruling landing in the next several months on ~75%-of-revenue North America (-). The permanent, unhedgeable Taiwan cross-strait tail (-). Rising AI-concentration (~61% of mix) on one capacity cycle (-).
Net: A quality franchise trading at a discount to lower-quality peers with an excellent catalyst slate, but carrying a genuine binary regulatory event and a structural geopolitical tail -> 7.
Data sourced from Daloopa (company_id 911). Market/consensus per FMP (2026-06-25); regulatory detail per USITC and public reporting.