TSM at a glance
The past year
Chart+37.2% $398.63, close Nov 13
- From the highest close
- −17.9%
- Vs the 200-day average
- −1.1%
Short interest
History32.2M shares
- Since the last report
- +9.9%
- Days to cover
- 2.6 days
- Settled
- Oct 30, 2026
Insider buying, 7 days
FilingsNone filed this week
- Form 4s filed
- Nov 10 – Nov 16
Top investors
Funds16 of 37 funds hold it
- David Tepper
- 10.2%
- Chase Coleman
- 9.7%
- Philippe Laffont
- 8.8%
- Latest 13F filings
- 9 added · 6 cut
Understand the Business
- Chip companies pay TSMC to fabricate wafers rather than own cutting-edge fabs; 7-nanometer and finer processes were 77% of 2Q26 wafer revenue, where performance-sensitive designs carry most value.↗
- Scale is part of the profit machine: in 2025 TSMC manufactured 12,682 products for 534 customers, helping spread process learning and factory costs across many designs.↗
Why this business is different
- TSMC is a manufacturing platform: it supports a global ecosystem with leading process technologies and design-enablement tools, making customers’ chip designs harder to move once built around its flow.↗
- The drawback is geography: management expected overseas fab ramp-up to dilute gross margin 2–3 percentage points, so trusted local capacity may earn less than Taiwan capacity.↗
What could matter next
Why the Stock Moved
- July 2025 · AI demand broadens: Expectations shifted toward artificial intelligence and high-performance computing as a larger, more durable load on leading-edge fabs. TSMC said 2Q25 was supported by robust artificial-intelligence and high-performance-computing demand. That raised future revenue and utilization confidence.↗↗
- January 2025 · Overseas costs surface: Expectations shifted toward a clearer margin drag from non-Taiwan fabs. Management forecast 2%–3% gross-margin dilution as overseas fabs ramped. That reduced the earnings benefit from geographic diversification.↗↗
TSM price history
Loading recent earnings…
Investment thesis
What must happen next
Artificial-intelligence chip demand must keep filling premium leading-edge capacity rather than fading into another semiconductor inventory cycle for earnings to hold.StrengtheningMateriality: High
2Q25 was supported by robust artificial-intelligence and high-performance-computing demand; 2Q26 advanced technologies reached 77% of wafer revenue. Sustained demand raises utilization and mix, which can turn expensive fabs into high incremental earnings.
Watch: Cloud customer order pushouts, weaker high-performance-computing platform mix, or falling advanced-node utilization would weaken this condition.
Supporting sources
Source notes are factual paraphrases; read the original disclosures below.
Our business in the second quarter was supported by continued robust AI and HPC-related demand.
Open source ↗shipments of 2-nanometer accounted for 3% of total wafer revenue; 3-nanometer accounted for 30%.
Open source ↗The 2-nanometer ramp must prove TSMC can keep customers moving to its newest process generation at meaningful volume.HoldingMateriality: High
2-nanometer shipments represented 3% of 2Q26 wafer revenue, while 3-nanometer represented 30%. New-node leadership extends the period in which TSMC can earn premium returns before older nodes face tougher pricing.
Watch: Slow 2-nanometer adoption, yield problems, or major customers delaying designs would challenge the durability of earnings growth.
Supporting sources
Source notes are factual paraphrases; read the original disclosures below.
shipments of 2-nanometer accounted for 3% of total wafer revenue; 3-nanometer accounted for 30%.
Open source ↗The broad customer and product base must remain a scale advantage for chip designers using TSMC’s manufacturing platform.HoldingMateriality: High
TSMC manufactured 12,682 products for 534 customers in 2025 and described a global ecosystem of customers and partners. Breadth improves learning, fills capacity across cycles, and makes design workflows harder to move away from.
Watch: Loss of a major design win or a narrowing customer base would reduce the scale and ecosystem advantage.
Supporting sources
Source notes are factual paraphrases; read the original disclosures below.
shipments of 2-nanometer accounted for 3% of total wafer revenue; 3-nanometer accounted for 30%.
Open source ↗Overseas fabs must improve customer trust without creating a permanently lower-margin manufacturing footprint than Taiwan-centered capacity over time.HoldingMateriality: High
Management forecast 2%–3% gross-margin dilution from overseas fabs as Kumamoto and Arizona ramped. Geographic diversification can help customers, but lower structural margins would reduce earnings on the same revenue.
Watch: Rising overseas dilution, delayed subsidies, or weaker overseas yields would make expansion less economically attractive.
Supporting sources
Source notes are factual paraphrases; read the original disclosures below.
we forecast 2% to 3% margin dilution impact from the ramp-up of our overseas fabs.
Open source ↗Capital spending must stay tied to advanced-process demand rather than outrunning customers during the artificial-intelligence buildout and weakening utilization.HoldingMateriality: Medium
TSMC planned a 2025 capital budget of $38 billion to $42 billion, with about 70% allocated to advanced process technologies. Discipline protects utilization and prevents depreciation from overwhelming profits if demand normalizes.
Watch: Large capacity additions followed by lower guidance or inventory digestion would signal weaker cash economics.
Supporting sources
Source notes are factual paraphrases; read the original disclosures below.
we forecast 2% to 3% margin dilution impact from the ramp-up of our overseas fabs.
Open source ↗
Full investment case & evidence
Loading the original research archive…
Original research
Artificial-intelligence chip demand must keep filling premium leading-edge capacity rather than fading into another semiconductor inventory cycle for earnings to hold. Sustained demand raises utilization and mix, which can turn expensive fabs into high incremental earnings.
The 2-nanometer ramp must prove TSMC can keep customers moving to its newest process generation at meaningful volume. New-node leadership extends the period in which TSMC can earn premium returns before older nodes face tougher pricing.
The broad customer and product base must remain a scale advantage for chip designers using TSMC’s manufacturing platform. Breadth improves learning, fills capacity across cycles, and makes design workflows harder to move away from.
Overseas fabs must improve customer trust without creating a permanently lower-margin manufacturing footprint than Taiwan-centered capacity over time. Geographic diversification can help customers, but lower structural margins would reduce earnings on the same revenue.
Capital spending must stay tied to advanced-process demand rather than outrunning customers during the artificial-intelligence buildout and weakening utilization. Discipline protects utilization and prevents depreciation from overwhelming profits if demand normalizes.
All standing thesis claims
Original report summary
leading-edge chip designers keep outsourcing manufacturing to TSMC, artificial-intelligence demand keeps fabs full, and overseas expansion does not turn strategic capacity into lower-return growth.
Since this case
What has happened since Sep 11, 2026
The case above was written from evidence through Sep 11, 2026 and stays as written. This part is read from SEC filings and closing prices each time the page is built, so it moves as new reports are filed.
- Results
- The financial statements could not be read just now.
- Share price
- The price since this case could not be read just now.
- Filings
- The filings since this case could not be read just now.
- Insiders
- Insider filings since this case could not be read just now.
- Conditions
- Sharp drawdown (Nov 5, 2026) crossed since Sep 11, 2026. What has followed each across the market is in the Thesis-Break panel.
Bull case and bear case
Bull case and bear case for Taiwan Semiconductor Manufacturing Co Ltd
Where the evidence on TSMC stands: Strengthening on the latest evidence (Sep 11, 2026). Of its 5 thesis points, 1 strengthening and 4 holding.
Bull case: what the evidence supports
Artificial-intelligence chip demand must keep filling premium leading-edge capacity rather than fading into another semiconductor inventory cycle for earnings to hold.
The 2-nanometer ramp must prove TSMC can keep customers moving to its newest process generation at meaningful volume.
The broad customer and product base must remain a scale advantage for chip designers using TSMC’s manufacturing platform.
Overseas fabs must improve customer trust without creating a permanently lower-margin manufacturing footprint than Taiwan-centered capacity over time.
Capital spending must stay tied to advanced-process demand rather than outrunning customers during the artificial-intelligence buildout and weakening utilization.
Bear case: what would count against it
Cloud customer order pushouts, weaker high-performance-computing platform mix, or falling advanced-node utilization would weaken this condition.
Slow 2-nanometer adoption, yield problems, or major customers delaying designs would challenge the durability of earnings growth.
Loss of a major design win or a narrowing customer base would reduce the scale and ecosystem advantage.
Company research
Research on TSMC
- Reporting · Taiwan Semiconductor Manufacturing Company Limited (NYSE: TSM) N2 Ramp Trims Margin
Third-quarter revenue accelerated as advanced chips gained share, while the first N2 production volume cut gross margin by 1.3 points.
Thesis-Break Engine
What has broken in TSMC’s thesis
Each break OptimistFi flagged on TSMC, dated — with what has historically followed the same kind of break across the market, not a forecast for TSMC.
- Sharp drawdowncondition set before this filing
Across all companies, this kind of break has happened 15,428 times before. Those stocks beat comparable companies by a median of 1% over the next six months, and 33% of them trailed it by 10 points or more.
Several kinds of break have stacked on TSMC at once. A cluster is a stronger read than any one signal alone.
MARKET + FINANCIAL SNAPSHOT
The numbers behind the thesis
What Changed
- SupportiveJuly revenue extended the strong run-rate
TSMC reported July 2026 revenue of NT$467.58 billion, up 5.6% month over month and 44.7% year over year, reinforcing that demand remained strong after 2Q26.
- Supportive2Q26 combined demand, margin, and new-node evidence
2Q26 revenue reached NT$1,270.38 billion and gross margin was 67.7%, while 2-nanometer entered wafer revenue; the combination supports the view that leading-edge demand is absorbing capacity.
Risks / falsifiers
What could break the thesis
- Cloud customer order pushouts, weaker high-performance-computing platform mix, or falling advanced-node utilization would weaken this condition.
- Slow 2-nanometer adoption, yield problems, or major customers delaying designs would challenge the durability of earnings growth.
- Loss of a major design win or a narrowing customer base would reduce the scale and ecosystem advantage.
- Rising overseas dilution, delayed subsidies, or weaker overseas yields would make expansion less economically attractive.
- Large capacity additions followed by lower guidance or inventory digestion would signal weaker cash economics.
TSM news
Loading headlines…
Financial statements
TSM income statement
Annual filed results · USD · Oldest to newest. Margins are calculated from the reported figures.
TSM: quick answers
- What is TSM's short interest?
- As of October 30, 2026, 32.2M TSM shares were sold short, up 9.9% from the count before. At the stock's average daily volume that is 2.6 days to cover. FINRA publishes the count twice a month. TSM short interest history
- Which well-known investors own TSM?
- 16 of the 37 well-known funds tracked here reported TSM shares in their latest 13F filings. By share of the fund's portfolio: Appaloosa (David Tepper), 10.2%; Tiger Global Management (Chase Coleman), 9.7%; Coatue Management (Philippe Laffont), 8.8%. Fund portfolios
TSM in well-known investors’ portfolios
16 of 37 funds reported TSM shares in their latest 13F filing. A 13F shows quarter-end positions and arrives up to 45 days later.
| Fund | Of its portfolio | Value | Change | Quarter |
|---|---|---|---|---|
| David TepperAppaloosa | 10.2% | $788M | +24.3% | Q2 2026 |
| Chase ColemanTiger Global Management | 9.7% | $2.3B | −12.3% | Q2 2026 |
| Philippe LaffontCoatue Management | 8.8% | $4.3B | −4.0% | Q2 2026 |
| Brad GerstnerAltimeter Capital | 6.9% | $683M | +4.7% | Q2 2026 |
| Leopold AschenbrennerSituational Awareness | 6.3% | $1.3B | +11,714% | Q2 2026 |
| Stanley DruckenmillerDuquesne Family Office | 5.4% | $282M | +19.1% | Q2 2026 |
| Dan LoebThird Point | 4.7% | $220M | +67.3% | Q2 2026 |
| Andreas HalvorsenViking Global Investors | 4.3% | $1.5B | −29.1% | Q2 2026 |
| Terry SmithFundsmith | 4.0% | $548M | New | Q2 2026 |
| Maverick Capital | 3.0% | $340M | −42.5% | Q2 2026 |
| George SorosSoros Fund Management | 2.9% | $250M | +5.1% | Q3 2026 |
| Whale Rock Capital | 2.5% | $281M | +4.3% | Q3 2026 |
The 12 with the largest share of their portfolio in TSM are shown. Every fund’s portfolio

