Management Quality -- 8.5/10
C.C. Wei (CEO since 2018) and Wendell Huang (CFO) run TSMC with Swiss-watch precision.
Capex has been delivered within guided ranges every year. Revenue has exceeded guidance
midpoint in 7 of 8 quarters. N3 yield ramp was ahead of schedule. Arizona Fab 1 achieved
yield comparable to Taiwan within months of volume production. Full-year 2025 revenue
guidance was raised three times -- from mid-20s% to ~30% to mid-30s% growth -- reflecting
management conservative-then-raise approach. The only minor knock: overseas fab dilution
of 2-3% is a known margin headwind that management is transparent about but cannot fully
offset near-term.
Weight: 20%
CEO Tenure
Since 2018
C.C. Wei -- 7+ years of consistent execution
Guidance Beat Rate
7/8 Quarters
Revenue at or above guidance midpoint
Capex Delivery
Within Range
FY2023, FY2024, FY2025 all within guided range
FY2025 Guidance Raises
Raised 3x
Mid-20s% to ~30% to mid-30s% growth
Capex guidance delivery
| Year | Capex Guidance | Actual | Status |
|---|---|---|---|
| FY2023 | $28-32B | $30.5B | HIT |
| FY2024 | $28-32B (initial), raised to $30-32B | $29.8B | HIT |
| FY2025 | $38-42B (initial), narrowed to $40-42B | $40.9B | HIT |
Management consistently delivers capex within guided ranges. Wendell Huang (CFO): "A higher
level of capital expenditures is always correlated with higher growth opportunities in the
following years."
Revenue guidance accuracy
| Quarter | Guidance Midpoint | Actual | Status |
|---|---|---|---|
| Q1 2025 | $28.8B | $25.5B | Met (seasonality) |
| Q2 2025 | $32.4B | $30.1B | Beat in USD (FX) |
| Q3 2025 | $32.8B | $33.1B | Beat |
| Q4 2025 | $35.2B | $33.7B | Met (FX headwind) |
Q2 2025 and Q4 2025 actual USD revenue was impacted by significant NT dollar appreciation
(4.4% and higher), which mechanically reduced USD-reported figures. Actual NT revenue
consistently beat or met guidance. Full-year 2025 revenue guidance was raised three times:
from mid-20s% to ~30% to mid-30s% growth.
N3/N5 yield ramp execution
One of the Fastest Node Ramps in TSMC History
N3 ramped from 9% of wafer revenue in Q1 2024 to 28% in Q4 2025 -- one of the fastest
node ramps in TSMC history. Arizona Fab 1 achieved "yield comparable to our fabs in
Taiwan" within months of volume production. N3 dilution expected to catch up to corporate
average "sometimes in 2026" per Wendell Huang. N2 development is "right on track" with
"structural profitability better than N3."
N3 Ramp Speed
9% to 28%
In 8 quarters (Q1 24 to Q4 25)
Arizona Fab 1 Yield
Taiwan-Comparable
Achieved within months of volume production
N2 Status
Right on Track
Structural profitability better than N3
Strategic vision
C.C. Wei has made several forward-looking calls that proved prescient:
- CoWoS expansion: Committed early and aggressively to CoWoS capacity expansion when AI demand signals were first emerging -- now vindicated by near-universal AI chip bottlenecks
- Arizona investment: $165B commitment demonstrates long-term strategic thinking about geographic diversification
- Customer engagement depth: Capacity planning now includes direct engagement with "customers' customers" (hyperscalers), providing unprecedented demand visibility
- Technology licensing discipline: Disciplined on not entering JVs or technology licensing despite pressure -- protecting the moat
- Pricing power: "Reflecting our value is a continuous and ongoing process" (Wendell Huang). N2 pricing will be higher than N3: "You don't expect our N2 is the same price as N3, right?" (C.C. Wei)
Known headwind: overseas fab dilution
2-3% Gross Margin Dilution, Narrowing Ahead of Schedule
The only minor knock on management: overseas fab dilution starts at 2-3% of gross margin,
with CFO noting "inflation in cost and potential tariff-related cost increases" widening
later-stage dilution to 3-4%. However, management is transparent about this headwind,
dilution narrowed from 2-3% to "closer to 2%" in H2 2025 (ahead of schedule), and
premium pricing partially offsets costs. Long-term gross margin floor of "53% and higher"
appears deliberately conservative given recent 59-62% actuals.
Score rationale
8.5/10. C.C. Wei and Wendell Huang deliver world-class execution.
Capex has been within guided ranges for 3 consecutive years. Revenue guidance has been
met or exceeded in 7 of 8 quarters. The full-year 2025 revenue guidance was raised
three times, reflecting a conservative-then-raise approach that builds credibility.
N3 yield ramp was one of the fastest in TSMC history. Arizona Fab 1 achieved
Taiwan-comparable yield. Strategic capital allocation balances aggressive growth
investment ($40.9B capex) with healthy FCF generation (NT$1,003B) and rising dividends.
The score does not reach 9+ because: (a) overseas fab dilution of 2-3% is a real margin headwind that management cannot fully offset near-term, widening to 3-4% as more fabs come online; (b) the long-term gross margin floor guidance of "53% and higher" may prove a ceiling if overseas expansion costs escalate; and (c) while management execution has been excellent in the current upcycle, the FY2023 downturn (-9% revenue) showed that even the best management cannot fully insulate against semiconductor cyclicality.
The score does not reach 9+ because: (a) overseas fab dilution of 2-3% is a real margin headwind that management cannot fully offset near-term, widening to 3-4% as more fabs come online; (b) the long-term gross margin floor guidance of "53% and higher" may prove a ceiling if overseas expansion costs escalate; and (c) while management execution has been excellent in the current upcycle, the FY2023 downturn (-9% revenue) showed that even the best management cannot fully insulate against semiconductor cyclicality.
Data sourced from Daloopa, earnings call transcripts Q1 2024 - Q4 2025.