The unit that drives semiconductor economics is cost per wafer. Taiwan Semiconductor Manufacturing Company built its position on a cost structure that decades of Taiwan-based operations made possible, and since Trump returned to power in 2025, the company has announced $200 billion in U.S. manufacturing investment. That figure is now running directly into the cost model that built TSMC's margin profile.
The cost differential and where it sits
Building and operating leading-edge fabs in the United States costs more than in Taiwan. Construction, operating overhead, and utility pricing all run higher on American soil. TSMC's margins were built on the Taiwan cost base. Shifting production capacity to U.S. soil, at the pace the Trump administration is pushing for on AI chip manufacturing specifically, reprices every input in that model upward.
When the headline reads "hits TSMC's margins," the mechanism is straightforward: the same output produced at higher input cost compresses the spread. The $200 billion signals the depth of the commitment. It also signals the scale of the exposure. Capital deployed into a structurally higher-cost environment does not produce the same margin as the same capital deployed into the Taiwan base.
The policy logic and what it costs TSMC
Trump's push is for American-made AI chips. The administration's premise is that domestic production reduces U.S. reliance on overseas semiconductor supply chains. For TSMC, meeting that push means building manufacturing capacity where the operating cost gap between the United States and Taiwan is a constant pressure, not a temporary adjustment.
The open question is where that gap lands: in TSMC's own margin structure, in higher prices passed to chip customers, or offset by some form of policy support. The $200 billion commitment, announced since Trump returned to power in 2025, makes that allocation the number the market is watching.