Apple's plan to use TSMC plants in the US and possibly Europe, a PR victory, buries the truth: capacity will be tiny and the chips will use an older 5nm process
Tim Culpan / Bloomberg : Tweets: @pstasiatech Tweets: Paul Triolo / @pstasiatech : Some commentators correctly point out that the new onshoring effort in the US under the CHIPS Act that has driven TSMC to site a facility in Arizona, will not be able in the short-term to make much of a dent in Taiwan dominance of high-end manufacturing. https://www.bloomberg.com/...
Context & Ripple Effects
TSMC's Arizona plant exists because the CHIPS Act pulled it there, and Tim Cook's celebration of it was pitched as a step toward US chip independence. The reporting underneath cuts against the ceremony: commentators like Paul Triolo note the facility will run an older 5nm process at tiny capacity, doing little to dent Taiwan's grip on high-end manufacturing. The project had already shown strain before this — TSMC and Intel struggled to recruit US workers to build the fabs, which began in 2021 with production targeted for 2024.
What came after confirms the skepticism was structural, not a one-off: TSMC's expanded US deal deliberately dropped one advanced-node fab while adding packaging and R&D, and Pat Gelsinger later argued even a $100B+ US investment pledge would not restore American leadership because R&D stays in Taiwan. Apple kept leaning in anyway, committing to billions of US-sourced chips and 100M+ units from Arizona by 2026.
First-order effects
- Apple gets a headline-friendly US supply story while its leading-edge silicon keeps coming from Taiwan — the Arizona fab covers a sliver of demand at a trailing node.
- TSMC contains its exposure: the US site handles older-process volume under political pressure, without moving the frontier out of Taiwan.
Second-order effects
- The gap between announcement and capability forces other players to calibrate expectations — Gelsinger's public dismissal of the $100B+ pledge frames Intel's own US buildout as the more serious onshoring bet.
- Packaging and R&D additions become the bargaining currency: TSMC trades visible US investment for keeping advanced-node fabs at home, shaping what future CHIPS Act deals look like.
Third-order effects
- If the pattern holds, US onshoring produces trailing-edge capacity and support functions while the semiconductor frontier stays concentrated in Taiwan — leaving supply-chain resilience claims resting on nodes generations behind state of the art.
- Buyers like Apple end up anchoring their reshoring commitments to capacity that cannot substitute for Taiwan in a crisis, meaning geopolitical risk pricing shifts to packaging, R&D location, and node vintage rather than headline fab counts.
The trend: CHIPS Act-era onshoring is delivering politically valuable but technologically lagging US capacity, leaving Taiwan's dominance of leading-edge manufacturing intact even as customers commit to buying American.