| Metric | Value | Note |
|---|---|---|
| Price (NYSE: TSM) | $467.67 (+1.20%) | AH $467.96; broke 52-wk high $476.79 intraday 6/22 |
| Returns | +15.6% 1M · +38% 3M · +53.9% YTD · +122% 1Y | Golden cross confirmed; uptrend intact |
| Market cap | $2.426T | 5.19B ADS; Asia's largest tech company |
| Forward P/E | 28.3x (NTM) | TTM P/E 39.6x; EPS TTM $11.82 |
| Gross / Net margin (TTM) | 62.1% / 47.3% | ROE 37.5%; debt/equity 18.6% |
| Revenue TTM | $130.5B | FY25 was $122.2B (+38.5% YoY) |
| Dividend yield | 0.81% | Beta 1.38 |
Price and stat data from CNBC quote page (TSM), June 23, 2026.
Q1 2026 (reported April 16): Revenue NT$1.134T / $35.9B, +40.6% YoY and +6.4% QoQ — a fourth consecutive record quarter. Gross margin 66.2%, operating margin 58.1%, net margin 50.5%. Advanced nodes (7nm and below) were ~74–75% of wafer revenue; sub-3nm was 25%. High-performance computing (AI + 5G) was 61% of revenue. CEO C.C. Wei: "AI-related demand continues to be extremely robust." (CNBC, Manufacturing Dive)
Q2 2026 guidance (the number that matters): Revenue $39.0B–$40.2B (~10% sequential), gross margin 65.5–67.5%, operating margin 56.5–58.5%. Full-year 2026 revenue growth guided above 30% in USD terms, with capex pushed to the high end of the $52–56B range (~$56B). (Manufacturing Dive)
The latest real-time tell — May monthly revenue: TSMC reported May 2026 net revenue of ~NT$416.98B, up ~30% YoY (filed June 10). Monthly prints are the freshest, least-massaged demand signal the company gives, and they are still running at +30% — directly on top of the +40% Q1. Counterpoint's William Li summed up the 2026 setup: "Demand still significantly outpaces supply… this sold-out environment will remain a defining characteristic." (May 2026 6-K via StockTitan)
This is the part of the thesis that compounds. Per sell-side channel checks circulating June 3–4, TSMC set 2nm (N2) wafer pricing at $30,000 per wafer for the early-2027 ramp — versus $20,000 for 3nm and $18,000 for 5nm. That is a 50% per-wafer increase on a single node transition, the steepest in two decades. Because early-node yields (~65–70%) produce ~25% fewer good die, the effective per-die cost rises ~80%.
Crucially, TSMC holds the pen on that price because there is no credible leading-edge alternative through end-2027. NVIDIA's Vera Rubin (2027), AMD's MI400, Broadcom's next-gen hyperscaler ASIC, Apple's A20, and Qualcomm's flagship are all on N2. They pay the new price or they lose their product cycle. The consensus margin model still assumes a more measured ~30% wafer increase — meaning the actual 50% print implies gross-margin expansion running ~2–3 points above consensus through 2027–2028. (Phemex Academy 2nm pricing breakdown)
Consequence #1: TSMC captures a structurally rising share of the AI compute dollar — it is the one node in the stack with monopoly pricing. Consequence #2: the wafer cost is the largest single input to cost-per-token; TSMC's pricing power effectively sets the floor under the entire AI cost curve, advantaging the largest-capex hyperscalers and tightening the screws on price-sensitive players (AMD, Qualcomm).
The single biggest overhang on sentiment is the U.S. foundry push. Intel began production of its 18A-P node (June 16), has reportedly struck an arrangement with Apple, and President Trump has publicly pushed Apple, NVIDIA and Tesla toward domestic Intel manufacturing. Intel is up ~200%+ YTD on this narrative (INTC ~$141, +5% on June 22). (CNBC: Intel 18A-P)
Why I think it's a sentiment story, not a 2026–2027 earnings story: NVIDIA tested Intel's 18A and did not commit; the leading-edge logic cohort re-upped on TSMC N2/N3. TSMC retains ~60–62% of total foundry revenue and 90%+ of leading-edge logic, and that doesn't change inside the current product cycle. Morningstar (wide-moat, 4-star, $428 fair value) called the Intel-share fear "overblown," noting TSMC is aggressively expanding its own advanced packaging — reinforced by the new 10-year TSMC–Amkor Arizona packaging pact (June 16). (Morningstar)
The real risk in this narrative is multi-year share compression (bulls' worst case: 70%→65% by 2028 if Samsung and Intel both execute) — not a 2026–2027 numbers problem. I am pricing it as a tail, not a base.
At 28x forward earnings for a business compounding revenue 30%+, with 66% gross margins, 47% net margins, 37% ROE and monopoly pricing on the most important node in technology, the PEG is below 1. Compare to its own customers riding TSMC-manufactured demand: NVDA, AMD (~99x forward), and AVGO all carry richer or comparable multiples on more cyclical, more competitive economics. You are paying less for the toll booth than for the cars driving through it.
Insider behavior corroborates: a cluster of TSMC executives were net buyers in early June (filings dated June 5–9; TipRanks headline "TSMC Rides Nvidia, Apple Deals as Insider Buys Big"). Insiders buying into a 52-week high is rare and bullish. Street targets are migrating up — Susquehanna's Mehdi Hosseini raised his target to $575 from $500 (June 22, Positive). (TipRanks/TheFly)
| Scenario | Prob. | Trigger | Path |
|---|---|---|---|
| Bull | 35% | Q3 guide ≥ +8% seq AND FY26 guide raised above 30%; N2 pre-bookings confirm $30K pricing holds | Multiple holds 28–30x on rising EPS → $560–600 |
| Base | 45% | In-line mid-July print, Q2 ~$40B, GM ~66%, FY guide reiterated >30% | Grinds higher with earnings → $510–540 |
| Bear | 20% | Hot PCE → 10Y > 4.55% multiple compression, OR GM guide cut <65% / FY trimmed below 30%, OR a hard Intel-Apple commitment headline | Re-rate to ~24x / breakout fails → $400–420 |
The stock is at an all-time high after +15.6% in a month and runs straight into two binary events: Core PCE on Thursday June 25 (8:30am ET, est ~+0.37% m/m) and TSMC's own mid-July print (June monthly revenue ~July 10; Q2 earnings ~July 16–19). A hot PCE that drives the 10Y back above 4.55% compresses long-duration tech multiples across the board — semis sell off first regardless of TSMC's own fundamentals. Do not chase the breakout into that. The higher-conviction approach is to accumulate on the pullback.
HIGH CONVICTION (thesis) WATCH (on chasing >$477)
Bottom line (my opinion): TSM is the position you hold for the entire AI cycle, not trade. The only thing wrong with it today is the entry point, not the company. Let PCE and the chase-crowd hand you a better price in the $435–460 zone; if it never comes and it closes cleanly above $477 on volume, that breakout is real and worth a starter.
Analysis for informational purposes only; not investment advice. Levels and probabilities are the author's judgment based on the data cited above as of June 23, 2026.