Revenue is compounding at 41% year-over-year in U.S. dollars and Q3 2026 net income skyrocketed 68.7% on insatiable AI demand — yet TSM stock trades essentially flat with forward estimates because the market is pricing one question: does the AI boom permanently expand TSMC’s moat, or does a $64B capex burden create an eventual margin trap? Four analyst lenses, three scenarios, four time horizons.
Gray line = TSM's actual price into today ($487 high Jun ’26 → $267 52-week low → $420.22 now); colored paths = synthesized scenario midpoints forward, probability-weighted (base 50% · bull 25% · bear 25%). Log-linear, mid-year marks. Wall Street 12-month consensus ≈ $584 (range $440–$1,000, “Strong Buy” from absolute majority of analysts).
Those probabilities are a judgment call — so make them yours. Drag to set how likely the bear and bull cases are (base takes the remainder); the blended target below, the dotted line on the chart, and the prob-weighted row of the scenario cards all update live.
The same fundamentals support wildly different conclusions depending on which framework you trust. Each lens below is a synthesized expert perspective with its own 12-month target.
TSMC operates a pure monopoly at the bleeding edge. AI demand is described by management as insatiable, and full-year 2026 revenue is guided up ~44%. The successful ramp of the 2nm node (N2) extends their technological lead, driving unparalleled pricing power with hyperscalers. A forward P/E around 19.5x is incredibly cheap for compounding 25%+ EPS growth.
Even with tens of billions in capital expenditures, TSM throws off over $45B in trailing free cash flow and sits on a pristine balance sheet with over $100B in net cash. Operating margins expanded to 58.7%. The business model converts silicon supremacy directly into massive shareholder yield, buffering against cyclicality.
Foundry economics are brutal when demand turns. TSMC is deploying up to $64B in future capital expenditures to satisfy the AI boom. If cloud hyperscaler spending cools, that massive capacity buildout becomes crushing depreciation, collapsing gross margins from the mid-60s down to the 40s. Plus, persistent China-Taiwan geopolitical risk demands a structural discount.
Intel's foundry ambitions are retreating, and Samsung continues to struggle with yields on advanced nodes. TSM isn't just winning market share; they are effectively the entire market for advanced compute. With 2nm production locking in hyperscalers for the next half-decade, their moat has never been wider or deeper.
What the sell-side expects over the next year. Bars are sorted low to high; the dashed line is today's $420.22 — note how nearly every target sits comfortably above it.
Sell-side 12-month targets — a selection of top firms covering TSM; the full consensus is ≈ $584, about +39% above today, with an overwhelming "Strong Buy" skew. The dashed line marks today's $420.22: literally every target currently rests above the live price, illustrating a profound disconnect between the fundamental AI momentum Wall Street projects and the cyclical/geopolitical anxiety the wider market is pricing. Firms, ratings, and targets illustrative.
Synthesized scenario midpoints (mid-year). Returns shown vs. today's $420.22. These are illustrative frameworks, not predictions with certainty — five-year outcomes hinge entirely on the durability of the AI spending cycle and geopolitics.
Where the money actually goes. TSM generates phenomenal free cash flow, but the bear thesis fixates entirely on the rapidly rising capex bar.
TSMC's extreme capital requirements in one view: revenue (sky) has surged aggressively alongside the AI boom, delivering over $45B in free cash flow (olive) this past year. However, maintaining Moore's Law and fulfilling AI demand requires jaw-dropping absolute capital expenditures (clay) — climbing toward $50B+ annually. The bear worry is that if revenue growth ever stalls, that massive capex bar transforms directly into margin-crushing depreciation. Debt (slate) remains structurally negligible. Figures illustrative based on USD approximations; 2026 represents forward 12-month projections.
The price targets are a function of an EPS estimate multiplied by an exit multiple. Here's the earnings ladder the scenarios are built on.
Adjusted EPS (per US ADR) — gray = reported, olive = forward base-case estimates assuming growth normalizes from its current hyper-growth phase toward a steady ~10–13% CAGR. The base case's ~$31 of 2031 EPS assigned a mature ~22× exit multiple is exactly what dictates the ~$680 base-case target. If the AI curve steepens, EPS pushes $40+ (bull case); if the hardware cycle busts, EPS plateaus mid-decade.
Q3 FY26, year-over-year — read these against a stock that recently shed 13%+ from its highs.
Every line is accelerating dramatically. Q3'26 revenue jumped 41% while EPS and Net Income soared nearly 70%, reflecting immense operating leverage. High-Performance Computing (HPC) and the advanced 3nm/2nm platforms (frontier segments, clay) are compounding at blistering rates, shifting the company's entire center of gravity to the bleeding edge. Wait times for capacity are stretching out years.
The entire valuation argument compresses into one core debate: is TSMC a permanent toll road with unassailable pricing power, or a cyclical manufacturer exposed to a vicious depreciation trap?
Where each risk sits, not just how big it is. The hot upper-right corner — likely and high-impact — is what prevents the multiple from expanding further; geopolitical tension and US/China export controls dominate the board.
Cross-strait tensions with China demand a permanent structural discount. Any outright blockade or conflict halts global tech supply chains instantly.
Building $50B+ in annual capacity for an AI cycle that could mature; idle advanced fabs generate crushing depreciation expenses.
Increasingly strict US regulations preventing TSMC from supplying advanced nodes to the Chinese market, capping global TAM.
Competitors eventually solve their yield issues, introducing price competition to the currently monopolized advanced foundry market.
A severe seismic event in Taiwan that breaches fab infrastructure, taking critical capacity offline for months.
EUV machines require enormous power; Taiwan's energy grid occasionally struggles, requiring TSMC to utilize backup power at higher costs.
New fabs in Arizona and Germany structurally carry lower gross margins and face difficult union/labor environments compared to Taiwan.
A mature smartphone upgrade cycle limits growth in the traditional mobile segment, though AI fully offsets this currently.
Chip manufacturing requires massive pure water usage. Severe droughts in Taiwan have historically required water trucking.
Hover the dotted terms in the metrics, or scan the desk's working definitions here.