According to english.cw.com.tw, Taiwan Semiconductor Manufacturing Company (TSMC) has announced a $100 billion investment in its Phoenix, Arizona campus—bringing its total planned U.S. commitment to $265 billion.
The Arizona Shockwave: Surprising the Market
The new commitment will fund at least four additional advanced fabrication facilities, expanding TSMC’s planned U.S. footprint to 12 facilities in total. These fabs will produce 2-nanometer (2nm) and below silicon, alongside advanced packaging capabilities such as CoWoS.
Building and operating these cutting-edge facilities in the United States is significantly more capital-intensive than in Taiwan—a factor contributing to Wall Street’s near-term caution. As Nobunaga Chai, senior industry analyst and deputy managing director of Cloud Express’ Institute of Political Economy and Industrial Research, observed:
“The sharp increase in this year’s capital expenditure and the additional US$100 billion to be invested in Arizona was a surprise to many analysts. There will be dilutions to the company’s gross profit margin for sure.” — Nobunaga Chai, senior industry analyst and deputy managing director, Cloud Express’ Institute of Political Economy and Industrial Research
The Domino Effect on Taiwan’s AI Supply Chain
TSMC’s expansion is driven by pressure from major U.S. tech buyers—and reportedly from the Trump administration—to secure leading-edge chip production domestically. According to an anonymous semiconductor industry analyst cited in the report, 2nm and below manufacturing in the U.S. is now a strategic imperative.
This has triggered a coordinated on-shoring wave among key Taiwanese suppliers. Foxconn, Quanta Computer, Wistron, Wiwynn, and TSMC’s critical materials and equipment suppliers are all expected to accelerate U.S. capital investments—co-locating near the Phoenix hub to enable end-to-end AI hardware delivery, from wafer start to fully packaged server rack.
Technology Pipeline: Entering the “Angstrom Era”
To sustain its multi-year U.S. buildout, TSMC unveiled its sub-2nm roadmap—including three nodes with precise timelines and innovations. The A14 (1.4nm) node enters risk production in 2027 and volume production in 2028. The A13 (1.3nm) node reaches mass production in 2029, while the A12 (1.2nm) node—featuring backside power delivery—also hits mass production in 2029.
The A12 node is specifically engineered for next-generation cloud AI servers, eliminating voltage drop and optimizing for extreme power demands. Each node delivers measurable gains: 15% speed boost and >20% logic density increase for A14; 6% area reduction for A13; and full backward compatibility across the series.
Massive Q2 Results Highlight a Two-Speed Market
Despite short-term market volatility—including a 7.3% drop in TSMC’s Taipei share price following the announcement—Q2 financial results underscored extraordinary momentum. Quarterly revenue reached $40.2 billion, net profit totaled NT$706.56 billion ($21.99 billion), and gross margin stood at 67.7%.
TSMC also raised its full-year capital expenditure guidance to $60 billion to $64 billion and lifted its full-year revenue growth forecast to slightly above 40% in U.S. dollar terms. Retail investor response was immediate: Taiwan’s odd-lot trading volume surged to 20.58 million shares—a 12-fold increase over the prior session.
These figures reflect a bifurcated semiconductor landscape: consumer electronics markets remain cautious, while demand for AI and High-Performance Computing (HPC) continues to outpace supply—locking TSMC into a multi-year deficit cycle.
Source: english.cw.com.tw
Compiled from international media by the SCI.AI editorial team.
