The result lands at a moment when some investors had begun questioning whether the industry’s breakneck pace of spending could be sustained. TSMC’s numbers suggest it can, at least for now.

Advanced chips are doing the heavy lifting. Processors measuring 7-nanometers or smaller accounted for 77% of wafer revenue in the previous quarter, up from 74% across full-year 2025. High-performance computing, which encompasses AI accelerators and data centre silicon, made up 55% of Q4 revenue and 58% of TSMC’s annual turnover, growing 48% year-on-year.

The wider supply chain is reading the same signals. Nvidia posted record full-year revenue of $215.9 billion for fiscal 2026, with its data centre division up 75% year-on-year. Foxconn, Nvidia’s largest server manufacturer, also reported a 30% rise in first-quarter revenue.

TSMC’s capital expenditure guidance tells its own story. The company expects to spend between $52 billion and $56 billion in 2026, up more than a third on last year, and has guided for full-year revenue growth of close to 30%. Chairman C.C. Wei has been characteristically direct about the rationale: AI demand looks “endless” and TSMC intends to be ready for it.

That readiness increasingly runs through the United States. TSMC is pressing ahead with a second Arizona fab, pulling forward its production schedule to the second half of 2027, while a third is already under construction. A fourth facility is in the permitting stage.

The Q1 beat will ease some nerves across the AI hardware ecosystem. For operators and infrastructure providers watching where capacity investment flows next, TSMC’s trajectory continues to point firmly upward.

RELATED STORIES

Taiwan positions itself as Washington’s AI partner as TSMC earnings underline strength of chip boom

AI chip demand continues to strain big tech supply chains