The Rationale That Repeats.
Three patterns show up across TAIWAN SEMICONDUCTOR MANUFACTURING's deal book — what the team buys, how it pays, and how it integrates. The patterns are the throughline; the deals below are the evidence.
01
Acquisition criteria
A foundry that grows organically, not by deal-making.
TSMC is a pure-play foundry whose scale comes overwhelmingly from building its own fabs, not from buying rivals. Genuine control acquisitions number roughly six across 25+ years of SEC filings, and the two largest (Worldwide Semiconductor and TSMC-Acer, both June 2000) were opportunistic capacity grabs during a demand spike, not a repeatable playbook.
02
Capital deployment
When TSMC does buy, it is buying capacity or an adjacent capability.
The foundry deals (WSMC, TSMC-Acer, the WaferTech buy-in) added installed wafer capacity fast; the smaller deals bought capabilities that feed the core foundry, namely SoC design services (GUC, 2003) and CMOS image-sensor back-end and wafer-level packaging (VisEra, 2015; Xintec, 2007). TSMC rarely enters a business it isn't already adjacent to.
03
Integration approach
Buy in, then let float.
Several holdings began as joint ventures TSMC bought into (WaferTech, VisEra) or majority stakes it later let dilute through IPOs (GUC to ~34.8%, Xintec to ~41.0%). TSMC also tends to raise minority stakes rather than acquire outright, for example lifting SSMC to 38.8% in 2006 and adding to Vanguard International Semiconductor (~27.6%) in 2024, treating equity positions as supply-chain and capacity tools rather than takeover targets.