Taiwan Semiconductor Manufacturing Company Limited
TSM deserves a BUY because AI-led demand is pushing earnings power materially higher, management is executing against that demand with credible guidance and capacity adds, and the stock still trades at 32.85x FY2025 EPS despite revenue now guided to grow slightly above 40% in FY2026.
TSM is being valued like a very good semiconductor company when it should be valued as the central manufacturing asset behind the AI compute buildout. The core fact is simple: management is guiding FY2026 revenue growth slightly above 40% YoY after already delivering FY2025 revenue of $121.3B, up 28.9%, and the demand signal inside that growth is high quality. In Q2 2026, HPC reached 66% of sales and grew 20% QoQ. That mix shift matters because HPC and AI demand are tied to leading-edge nodes and advanced packaging where TSM's moat is deepest.
The market is right to worry about cost. TSM is spending heavily, with FY2026 CapEx raised to $60B-$64B, and management has been explicit that the 2nm ramp will dilute gross margin by 3-4 points in H2 2026 while overseas fabs dilute another 2%-4%. Those are real offsets. But they need to be judged against what TSM is buying with that spend: more 3nm fabs, conversion of 5nm tools to 3nm, added 2nm-and-below capacity in Arizona, and continued packaging expansion into a market management still describes as supply constrained. This is not defensive capex. It is share-protecting, moat-deepening capex into visible demand.
The stock price matters here because it shapes the setup. At $450.61, shares are only 5.9% below the 52-week high of $479.00 and 17.6% above the 200-day average, so this is not an early-cycle bargain. That limits conviction and means the next 12 months must be driven by delivered earnings, not just enthusiasm for AI. Still, the valuation is not stretched relative to growth. P/E is 32.85x versus a 43.4x peer average, while EPS grew 36.9% in FY2025 and management just raised the 2026 outlook.
The case depends on three things. First, AI demand must remain strong enough to keep leading-edge utilization high. Second, 2nm and packaging have to ramp without a lasting margin break. Third, management must keep executing at its recent standard. So far, it has done exactly that, with 8 straight EPS beats and Q2 revenue landing at the high end of guidance. That combination supports a BUY, but with medium rather than high conviction because valuation now requires continued delivery.
AI demand remains extremely robust, packaging constraints ease, 2nm ramps cleaner than feared, and the market rewards TSM with a premium closer to high-quality AI infrastructure peers.
FY2026 revenue grows slightly above 40%, advanced-node utilization stays high, and earnings expansion outweighs margin dilution enough to support the engine target.
2nm ramp dilution and overseas fab costs push margins below plan, AI packaging stays constrained, and the market refuses to pay above low-30s earnings multiples near cycle highs.
Quant call aligns with analyst consensus (Strong Buy: 6 buy / 1 hold / 0 sell).
Go for a business that any idiot can run — because sooner or later, any idiot probably will.
Forward earnings power is moving much faster than the trailing numbers imply. On FY2025 EPS of $10.50, the stock trades at 32.85x, but if management delivers slightly above 40% FY2026 revenue growth with Q3 operating margin guided at 56%-58%, the forward P/E drops materially against a business whose AI mix is still rising. The moat looks stronger, not weaker, over the next 3-5 years because leading-edge logic, packaging, and customer trust are getting harder to replicate as 2nm and A14 advance, though geopolitics and overseas fab dilution cap how much multiple expansion is realistic. AI is clearly a net tailwind here, not a disruption risk, because TSM is the manufacturing bottleneck behind many of the winners.
- FY2025 revenue grew 28.9% to $121.3B and EPS grew 36.9% to $10.50, showing operating leverage is already flowing through the model.
- HPC reached 66% of Q2 2026 sales and grew 20% QoQ, a direct sign that AI compute demand is now the core earnings engine.
- Management raised FY2026 CapEx to $60B-$64B and now expects FY2026 revenue growth slightly above 40% YoY, indicating demand visibility is strong enough to fund a much larger buildout.
- Q2 2026 revenue hit $40.2B, the high end of guidance, and gross margin reached 67.7%, ahead of guidance despite heavy node transitions.
- TSM has beaten EPS estimates in 8 straight quarters, with beats ranging from +3.4% to +12.7%, including +11.4% in July 2026, which supports management credibility on demand and cost control.
- Q3 2026 revenue guidance of $44.6B-$45.8B implies 12% sequential growth at the midpoint even after a very strong Q2, which is unusual scale for a company already at this size.
- P/E of 32.85 sits below the 43.4x peer average, giving investors a way to own the dominant advanced foundry franchise without paying the highest peer multiple.
- Management explicitly expects the 2nm ramp to dilute gross margin by 3-4 points in H2 2026, while overseas fabs still dilute margin by 2%-4%, so earnings growth will not flow through cleanly every quarter.
- CapEx of $60B-$64B against FCF of $9.1B means current free cash flow is under pressure from an aggressive build cycle, especially with an additional $100B Arizona investment announced.
- P/S of 16.66 is well above the 11.3 peer average, so valuation already embeds a premium for quality and AI leadership even if earnings multiples look more reasonable.
- The quant engine's peer set shows 3.8x wide P/E dispersion, which weakens confidence in multiple-based upside and raises the chance the market keeps a discounted framework on foundries versus fabless AI names.
| Metric | Value | Context |
|---|---|---|
| Price target | $578.94 | Ground-truth engine target, 28.5% above current $450.61, low confidence |
| FY2025 revenue | $121.3B | +28.9% YoY, strong scale growth into 2026 |
| FY2025 EPS | $10.50 | +36.9% YoY, outgrowing revenue |
| Operating margin | 45.7% | Flat versus prior FY despite heavy investment cycle |
| Net margin | 40.5% | Exceptional profitability for a capital-intensive manufacturer |
| Valuation | P/E 32.85x | P/S 16.66x | P/B 11.53x | P/E below 43.4x peer avg, but P/S above 11.3x peer avg |
| Balance sheet | Debt/Equity 0.17 | Debt/EBITDA 0.39 | Current ratio 2.46 | Low leverage and strong liquidity |
| Cash generation | FCF $9.1B | Positive FCF despite elevated investment needs |
| FY2026 guidance | Revenue growth slightly above 40% YoY | CapEx $60B-$64B | Raised outlook signals strong demand visibility |
| Q3 2026 guidance | Revenue $44.6B-$45.8B | Gross margin 65%-67% | Operating margin 56%-58% | Implies 12% sequential revenue growth at midpoint |
| Earnings track record | 8 straight EPS beats | Latest was 4.31 vs 3.87, a +11.4% beat |
| 52-week position | $450.61 vs $479.00 high | Shares are -5.9% below the high and 17.6% above the 200-day average |
The biggest risk is that the 3-4 point H2 2026 gross margin hit from the 2nm ramp lasts longer than planned while $60B-$64B of FY2026 CapEx stays elevated, compressing returns even if revenue remains strong.
The next catalyst is October 15, 2026 earnings, where investors need confirmation that Q3 revenue lands within $44.6B-$45.8B and that the FY2026 growth outlook of slightly above 40% holds.
Ownership looks institutionally sponsored but not crowded by insider control: the top 3 institutions hold 14.2%, insiders hold 0.0%, and the short base is light at 0.54% of float with 2.2 days-to-cover. That means there is little short-squeeze fuel, but also limited bearish conviction. Options positioning is supportive, with a positive GEX regime and a zero-gamma level of $429.44, about 4.83% below spot, which suggests dealer positioning is more likely to damp volatility than amplify it unless the stock breaks materially lower.
Up 20% QoQ, now the dominant driver as AI and data center demand offset weaker smartphone demand.
Weaker demand persists in consumer and other price-sensitive markets, according to management.
70%-80% of FY2026 CapEx is directed to advanced process, showing where management sees the economic center of demand.
Management said this revenue would more than triple in 2024; current AI-led HPC mix suggests that contribution has continued to scale since then.
TSMC posted strong Q2 results and guided double-digit Q3 growth while raising 2026 CapEx and FY26 revenue outlook.
Management has earned credibility. On the 2024 Q3 call, it guided Q4 2024 revenue to $26.1B-$26.9B, said AI demand was "extremely robust," forecast FY2024 growth close to 30%, and signaled 2025 CapEx would be higher than 2024. Since then, the company delivered FY2025 revenue of $121.3B, up 28.9%, EPS of $10.50, up 36.9%, and kept operating margin at 45.7% even through a heavy investment phase. The strategic direction has not changed, it has intensified: CoWoS remained tight, advanced-node demand strengthened, 2nm demand is ramping, and FY2026 CapEx is now $60B-$64B, roughly double the slightly above $30B FY2024 level management discussed two years ago. The beat pattern also supports trust, with 8 straight EPS beats from +3.4% to +12.7%.
Based on the current business outlook, we expect our third quarter revenue to be between $44.6 billion and $45.8 billion, which represents a 12% sequential increase or a 37% year-over-year increase at the midpoint.
Given the continued strong structural demand from our customers, including the newly emerging agentic AI market, we have decided to raise our full year 2026 capital budget to be between $60 billion and $64 billion as we continue to invest heavily to support our customers' growth.
We concluded our second quarter with revenue of $40.2 billion at the high end of our guidance in U.S. dollar terms, driven by strong demand for our leading-edge process technologies.
Supported by our robust technology differentiation and broad customer base, we now expect our full year 2026 revenue growth to be slightly above 40% year-over-year in U.S. dollar terms.
Having said that, AI related demand continues to be extremely robust. Thus, our conviction in the multi-year AI megatrend remains very high.
TSM's moat is process leadership plus manufacturing trust at scale. Customers need a foundry that can ramp 3nm, 2nm, and advanced packaging with high yield, and TSM is still the one spending $60B-$64B a year to stay ahead while serving a broad customer base. That moat should remain durable over the next 3-5 years because the cost, engineering talent, ecosystem integration, and learning-curve advantages required to match it are getting larger, not smaller.
Capital allocation is aggressive but rational for the demand backdrop. TSM generated $9.1B of FCF, spent $15.6B on capex, paid $4.9B in dividends, and did $0.0B in buybacks, which is the right order of priorities for a foundry in a major node transition. The raised $60B-$64B FY2026 capital budget and added $100B Arizona investment increase near-term cash strain, but they also reinforce that management is funding scarce assets where customer demand is strongest.
TSM rates a BUY with Medium conviction. The company is executing like the indispensable manufacturing backbone of the AI cycle, and the forward revenue and earnings path still looks better than the current valuation implies. The stock is not cheap on every metric and it is no longer washed out technically, so upside now depends on continued delivery against an elevated bar. A downgrade would require evidence that 2nm margin dilution is becoming structural or that the slightly above 40% FY2026 growth outlook is slipping. Until then, the numbers support staying long.
≣ Full research note tap to expand
Executive Summary
BUY, Medium conviction. Rated BUY, Medium conviction. TSM is no longer cheap on simple asset or sales multiples, but the market is still underpricing how much AI, 2nm, and advanced packaging can lift forward earnings power as management guides Q3 2026 revenue to $44.6B-$45.8B and FY2026 growth to slightly above 40% YoY. The setup is less forgiving with shares only 5.9% below the 52-week high, so this is a thesis on continued estimate expansion, not on multiple reversion alone.
Investment Thesis
TSM is being valued like a very good semiconductor company when it should be valued as the central manufacturing asset behind the AI compute buildout. The core fact is simple: management is guiding FY2026 revenue growth slightly above 40% YoY after already delivering FY2025 revenue of $121.3B, up 28.9%, and the demand signal inside that growth is high quality. In Q2 2026, HPC reached 66% of sales and grew 20% QoQ. That mix shift matters because HPC and AI demand are tied to leading-edge nodes and advanced packaging where TSM's moat is deepest.
The market is right to worry about cost. TSM is spending heavily, with FY2026 CapEx raised to $60B-$64B, and management has been explicit that the 2nm ramp will dilute gross margin by 3-4 points in H2 2026 while overseas fabs dilute another 2%-4%. Those are real offsets. But they need to be judged against what TSM is buying with that spend: more 3nm fabs, conversion of 5nm tools to 3nm, added 2nm-and-below capacity in Arizona, and continued packaging expansion into a market management still describes as supply constrained. This is not defensive capex. It is share-protecting, moat-deepening capex into visible demand.
The stock price matters here because it shapes the setup. At $450.61, shares are only 5.9% below the 52-week high of $479.00 and 17.6% above the 200-day average, so this is not an early-cycle bargain. That limits conviction and means the next 12 months must be driven by delivered earnings, not just enthusiasm for AI. Still, the valuation is not stretched relative to growth. P/E is 32.85x versus a 43.4x peer average, while EPS grew 36.9% in FY2025 and management just raised the 2026 outlook.
The case depends on three things. First, AI demand must remain strong enough to keep leading-edge utilization high. Second, 2nm and packaging have to ramp without a lasting margin break. Third, management must keep executing at its recent standard. So far, it has done exactly that, with 8 straight EPS beats and Q2 revenue landing at the high end of guidance. That combination supports a BUY, but with medium rather than high conviction because valuation now requires continued delivery.
Key Metrics
| Metric | Value | Context | ||
|---|---|---|---|---|
| Price target | $578.94 | Ground-truth engine target, 28.5% above current $450.61, low confidence | ||
| FY2025 revenue | $121.3B | +28.9% YoY, strong scale growth into 2026 | ||
| FY2025 EPS | $10.50 | +36.9% YoY, outgrowing revenue | ||
| Operating margin | 45.7% | Flat versus prior FY despite heavy investment cycle | ||
| Net margin | 40.5% | Exceptional profitability for a capital-intensive manufacturer | ||
| Valuation | P/E 32.85x | P/S 16.66x | P/B 11.53x | P/E below 43.4x peer avg, but P/S above 11.3x peer avg |
| Balance sheet | Debt/Equity 0.17 | Debt/EBITDA 0.39 | Current ratio 2.46 | Low leverage and strong liquidity |
| Cash generation | FCF $9.1B | Positive FCF despite elevated investment needs | ||
| FY2026 guidance | Revenue growth slightly above 40% YoY | CapEx $60B-$64B | Raised outlook signals strong demand visibility | |
| Q3 2026 guidance | Revenue $44.6B-$45.8B | Gross margin 65%-67% | Operating margin 56%-58% | Implies 12% sequential revenue growth at midpoint |
| Earnings track record | 8 straight EPS beats | Latest was 4.31 vs 3.87, a +11.4% beat | ||
| 52-week position | $450.61 vs $479.00 high | Shares are -5.9% below the high and 17.6% above the 200-day average |
Financial Strength
Financial strength is a major positive. Leverage is low at 0.17 debt/equity and 0.39 debt/EBITDA, liquidity is solid with a 2.46 current ratio, and profitability remains elite with 45.7% operating margin and 40.5% net margin. The main caveat is not solvency, it is cash conversion during a historic investment cycle: FCF of $9.1B is positive but modest relative to the scale of capacity expansion, so investors should expect periods where accounting earnings look much better than free cash flow while new fabs ramp.
Competitive Position
TSM's moat is process leadership plus manufacturing trust at scale. Customers need a foundry that can ramp 3nm, 2nm, and advanced packaging with high yield, and TSM is still the one spending $60B-$64B a year to stay ahead while serving a broad customer base. That moat should remain durable over the next 3-5 years because the cost, engineering talent, ecosystem integration, and learning-curve advantages required to match it are getting larger, not smaller.
Management & Guidance
Guidance is strong and, importantly, detailed. Management expects Q3 2026 revenue of $44.6B-$45.8B, with 65%-67% gross margin and 56%-58% operating margin, and raised the full-year outlook to slightly above 40% revenue growth with $60B-$64B of capex. Credibility is above average because Q2 revenue landed at the high end of guidance, gross margin beat guidance at 67.7%, and the company has posted 8 consecutive EPS beats. The one area to treat conservatively is margin timing, because management has already flagged a 3-4 point H2 hit from the 2nm ramp.
Segment Analysis
- HPC (66% of Q2 2026 revenue): Up 20% QoQ, now the dominant driver as AI and data center demand offset weaker smartphone demand.
- Smartphone and consumer-oriented end markets (Not disclosed in provided data): Weaker demand persists in consumer and other price-sensitive markets, according to management.
- Leading-edge process and advanced packaging (Not disclosed in provided data): 70%-80% of FY2026 CapEx is directed to advanced process, showing where management sees the economic center of demand.
- Server AI processors (Mid-teens of FY2024 revenue per historical call): Management said this revenue would more than triple in 2024; current AI-led HPC mix suggests that contribution has continued to scale since then.
Capital Allocation
Capital allocation is aggressive but rational for the demand backdrop. TSM generated $9.1B of FCF, spent $15.6B on capex, paid $4.9B in dividends, and did $0.0B in buybacks, which is the right order of priorities for a foundry in a major node transition. The raised $60B-$64B FY2026 capital budget and added $100B Arizona investment increase near-term cash strain, but they also reinforce that management is funding scarce assets where customer demand is strongest.
Management Execution & Track Record
Management has earned credibility. On the 2024 Q3 call, it guided Q4 2024 revenue to $26.1B-$26.9B, said AI demand was "extremely robust," forecast FY2024 growth close to 30%, and signaled 2025 CapEx would be higher than 2024. Since then, the company delivered FY2025 revenue of $121.3B, up 28.9%, EPS of $10.50, up 36.9%, and kept operating margin at 45.7% even through a heavy investment phase. The strategic direction has not changed, it has intensified: CoWoS remained tight, advanced-node demand strengthened, 2nm demand is ramping, and FY2026 CapEx is now $60B-$64B, roughly double the slightly above $30B FY2024 level management discussed two years ago. The beat pattern also supports trust, with 8 straight EPS beats from +3.4% to +12.7%.
Positioning & Flows
Ownership looks institutionally sponsored but not crowded by insider control: the top 3 institutions hold 14.2%, insiders hold 0.0%, and the short base is light at 0.54% of float with 2.2 days-to-cover. That means there is little short-squeeze fuel, but also limited bearish conviction. Options positioning is supportive, with a positive GEX regime and a zero-gamma level of $429.44, about 4.83% below spot, which suggests dealer positioning is more likely to damp volatility than amplify it unless the stock breaks materially lower.
Bull Case
- Q2 2026 revenue hit $40.2B, the high end of guidance, and gross margin reached 67.7%, ahead of guidance despite heavy node transitions.
- TSM has beaten EPS estimates in 8 straight quarters, with beats ranging from +3.4% to +12.7%, including +11.4% in July 2026, which supports management credibility on demand and cost control.
- Q3 2026 revenue guidance of $44.6B-$45.8B implies 12% sequential growth at the midpoint even after a very strong Q2, which is unusual scale for a company already at this size.
- P/E of 32.85 sits below the 43.4x peer average, giving investors a way to own the dominant advanced foundry franchise without paying the highest peer multiple.
Bear Case
- Management explicitly expects the 2nm ramp to dilute gross margin by 3-4 points in H2 2026, while overseas fabs still dilute margin by 2%-4%, so earnings growth will not flow through cleanly every quarter.
- CapEx of $60B-$64B against FCF of $9.1B means current free cash flow is under pressure from an aggressive build cycle, especially with an additional $100B Arizona investment announced.
- P/S of 16.66 is well above the 11.3 peer average, so valuation already embeds a premium for quality and AI leadership even if earnings multiples look more reasonable.
- The quant engine's peer set shows 3.8x wide P/E dispersion, which weakens confidence in multiple-based upside and raises the chance the market keeps a discounted framework on foundries versus fabless AI names.
Valuation & Price Target
Engine price target $578.94 (+28.5% vs current), Low confidence.
- P/E Multiple: $24934.6 (0% weight)
- PEG (growth-adjusted): $25157.45 (0% weight)
- P/S Multiple: $295.98 (20% weight)
- Quality-adjusted: $547.94 (20% weight)
- DCF: $130.36 (5% weight)
This analysis aligns with the quant baseline BUY, but the conviction stays Medium, not High, because the engine's $578.94 target carries Low confidence and leans partly on peer multiples despite 3.8x dispersion in the comp set. The key reason to still own it is not the peer framework, it is the company's own forward trajectory: FY2025 EPS of $10.50, Q3 2026 revenue guidance of $44.6B-$45.8B, and FY2026 growth slightly above 40% point to a much stronger earnings base than trailing valuation screens capture. Sell-side consensus is also bullish, but that is only a cross-check, not the basis of the call.
Risk Assessment
Risk score 22/100 (Low). The biggest risk is that the 3-4 point H2 2026 gross margin hit from the 2nm ramp lasts longer than planned while $60B-$64B of FY2026 CapEx stays elevated, compressing returns even if revenue remains strong.
The Bottom Line
TSM rates a BUY with Medium conviction. The company is executing like the indispensable manufacturing backbone of the AI cycle, and the forward revenue and earnings path still looks better than the current valuation implies. The stock is not cheap on every metric and it is no longer washed out technically, so upside now depends on continued delivery against an elevated bar. A downgrade would require evidence that 2nm margin dilution is becoming structural or that the slightly above 40% FY2026 growth outlook is slipping. Until then, the numbers support staying long.
Go deeper in the app
Live charts, options analytics, the full earnings-call breakdown, price alerts, and every other stock — updated in real time.
Get Tickr