Current Setup & Catalysts

Figures converted from TWD at historical FX rates — see data/company.json.fx_rates. Ratios, margins, and multiples are unitless and unchanged. April 2026 monthly revenue converted at ~31.59 NT$/US$; dividends converted at the May 2026 spot rate.

Current Setup & Catalysts

The setup is a crowded long inside a richly-priced compounder that just printed the cleanest quarter in its history (Q1 2026 GM 66.2%, USD revenue +40.6% YoY) and then walked into its first soft monthly print of the year (April revenue US$13.0B, +17.5% YoY but −1.1% MoM — the first sequential drop since October 2025). The stock sits 4% below its April all-time high (ADR US$404.35 on 2026-05-15), 26.5% YTD, 129% over twelve months, and the live debate is not whether TSMC's leading-edge monopoly is real — Q1 already proved it — but whether the 66% gross margin is the new regime or a one-quarter mix peak before N2 dilution, Arizona absorption, and a possible hyperscaler AI capex digestion compress the print. The next hard date that matters is Q2 2026 earnings (mid-July 2026): management's own guide (GM 65.5–67.5%) tells you they think the regime holds, but every bear catalyst — Apple multi-sourcing rumours, Samsung 2nm narrative, Section 232 follow-on, the Wei-Jen Lo case verdict — runs on a softer-dated clock that can re-rate the multiple inside a single news cycle.

1. Current Setup in One Page

2. What Changed in the Last 3–6 Months

The recent window is dense with events that re-rated both the regime case (margins, capacity, demand) and the risk case (IP, tariffs, customer multi-sourcing). The eight items below are the ones that still control today's setup.

The recent narrative arc is straightforward: from January through mid-April, the market priced TSMC as the consensus AI implementation trade (98% buy ratings; stock +33% YTD by mid-April). The Q1 print on April 16 confirmed that view — record revenue, GM above the upper guide, FY26 guide raised — and the stock made an all-time high (US$419.50) shortly after. From late April onward, three things shifted: (i) Apple/multi-sourcing chatter resurfaced (still rumour, not contract); (ii) the first soft monthly revenue print landed (April −1.1% MoM); and (iii) Taiwan moved to the center of the Trump-Xi conversation. What investors used to worry about (cycle, capex absorption, China decoupling) is now mostly priced as resolved; what they worry about now is regime durability — whether the 66% GM holds when N2 dilution + Arizona absorption + a possible AI capex digestion stack up in 2H26. The Wei-Jen Lo / Tokyo Electron IP cases sit in the background as a slow-burn moat test that none of the bullets above quantify.

3. What the Market Is Watching Now

The live debate has four tracks. Each item is a discrete data point with a verifiable expectation gap.

These four tracks share a single property: none of them resolve at Q2 earnings alone. Q2 confirms the GM regime; the customer-concentration and geopolitical tracks resolve over multiple quarters via court calendars, agency filings, and customer commentary. That asymmetry is the live setup — the binary "regime holds" data point lands in July; the structural "moat survives" data points keep landing through 2026–27.

4. Ranked Catalyst Timeline

Ranked by decision value to an institutional underwriter, not chronology. Where dates are not yet announced, the window is conservative based on TSMC's historical cadence.

5. Impact Matrix

The catalysts above sort into three tiers by what they actually resolve. The matrix below focuses on the items that update a durable thesis variable, not those that merely add news.

The matrix sorts cleanly: items 1–3 update the long-term thesis; items 4–6 update the near-term evidence path without re-rating the franchise unless extreme. The only catalyst that can flip the long-term thesis from "wide moat compounder" to "narrow moat at peak cycle" inside a single news cycle is item 2 — Samsung 2nm with a named HPC customer. Q2 earnings (item 1) re-rates the multiple but not the moat; hyperscaler capex (item 3) re-rates the earnings curve but not the franchise. Items 4–6 add noise and trim/raise the risk premium but do not change the underwriting baseline.

6. Next 90 Days

The next-90-day window is unusually busy for TSMC by historical standards because the Q2 earnings cadence collides with two monthly revenue releases, a tech-symposium aftermath, the dividend ex-date, the AGM, and the hyperscaler Q2 reporting cycle.

A PM should think about TSMC over the next 90 days as a guided run-up: two monthly prints, a known dividend ex-date, and then the Q2 print that decides the GM regime debate for the remainder of 2026. The window has more decision-relevant content than any 90 days in the prior twelve months.

7. What Would Change the View

Three observable signals would most change the investment debate over the next six months. First, Q2 FY2026 gross margin landing at the lower bound of the 65.5–67.5% guide with explicit walk-down commentary for Q3/Q4 would convert the "regime holds" call into "Q1 was the peak" — that single print would reject the 56%+ through-cycle anchor as a floor and put the multiple at risk of compressing toward the 10-year average (P/E ~22 vs current ~29). Second, a Samsung Foundry quarter that reverses the FY25 −3.9% trajectory and arrives with a publicly named HPC customer at 2nm — the single variable the Long-Term Thesis flags as load-bearing — would downgrade the wide moat to narrow inside one earnings cycle, irrespective of Q2's GM print. Third, two consecutive quarters of flat-to-down aggregate hyperscaler AI capex (MSFT+GOOG+META+AMZN+ORCL) coupled with TSMC capex productivity (Rev/Net PP&E) drifting from 1.2x toward 1.0x — the Bear's stated primary trigger — would activate the demand-digestion thesis with FY26 capex US$52–56B already pre-loading depreciation. None of these three has yet hit; the present setup is the calm before the data points that would test them. The first proof-point on path 1 is mid-July 2026; paths 2 and 3 are continuous through 2026–27 and resolve outside any single earnings cycle.