TSMC raises its 2026 capex projection from $52B-$56B to $60B-$64B and its revenue growth projection in USD from 30%+ to 40%+ YoY, citing “the AI megatrend”
Debby Wu /Bloomberg:
Context & Ripple Effects
TSMC began 2026 expecting $52B-$56B in capital spending and close to 30% USD revenue growth, already above analyst expectations. The new outlook materially lifts both targets, extending a run of upward revisions tied to strong AI demand through 2025.
The company’s latest results also show demand concentrated in advanced production: chips at 7nm or below made up 77% of Q2 wafer revenue, while TSMC is expanding advanced packaging capacity alongside wafer fabrication.
First-order effects
- TSMC will direct an additional $8B of planned 2026 investment toward capacity and supporting manufacturing infrastructure while resetting investor expectations for materially faster sales growth.
- The revised plan reinforces TSMC’s ability to serve current demand for advanced chips and packaging, while increasing its near-term exposure to the execution and cost of a larger build-out.
Second-order effects
- Equipment and advanced-packaging suppliers gain a stronger demand signal, although TSMC’s reported resistance to ASML price increases indicates it will seek to preserve leverage over the cost of that expansion.
- Customers relying on leading-edge manufacturing get a clearer indication that TSMC intends to add supply; rival foundries face greater pressure to match its technology and capacity investment rather than compete primarily on existing capacity.
Third-order effects
- If repeated upgrades persist, AI infrastructure demand is likely to keep shifting semiconductor value and capital intensity toward leading-edge fabrication and packaging, where scale and execution matter most.
- The scale of TSMC’s planned investment, including its stated US build-out, could further concentrate strategic chip supply in a small group of heavily capitalized manufacturers while making location and supply-chain policy more consequential.
The trend: This is another data point in the AI-led conversion of advanced semiconductor manufacturing and packaging from a cyclical capacity business into a sustained, capital-intensive strategic bottleneck.