The chipmaking giant TSMC plans to invest another $100 billion in its Arizona facilities, driven by robust demand from U.S. clients and the need to fend off aggressive competitors, Chief Financial Officer Wendell Huang stated this week.
The firm, having just reported better-than-expected sales and earnings for the three months ending in June and raising its outlook, continues to face difficulties in meeting all client orders. What TSMC refers to as the artificial intelligence megatrend will fuel its expansion for at least the coming years.
Huang remarked during a Friday interview, "We are doing everything we can, wherever we can, to expand to support the customers' growth."
TSMC is contending with increasing rivalry from Intel and Elon Musk's Terafab. Huang further said that TSMC aims to "demonstrate that we do not intend to leave any food on the table to anybody else."
TSMC, the primary chip supplier for Nvidia and Apple, announced the extra spending at its Thursday earnings call.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
According to Huang, the scale of this investment renders it the biggest foreign direct investment in U.S. history, and the timeline for completion will hinge on market conditions. TSMC, with one fabrication plant already running, plans to bring two additional fabs online before 2030 and will construct further facilities at its Arizona site.
CEO C.C. Wei said on Thursday that the additional $100 billion might lead to four more factories, possibly bringing TSMC's Arizona complex to a total of ten chipmaking plants plus two packaging sites.
TSMC's decision to consistently expand its U.S. investments is also fueled by geopolitical factors and a desire to be near major AI clients. North American clients such as Nvidia and AMD generated 78% of the firm's total revenue last quarter.
The top four U.S. AI operators and developers, spanning Meta Platforms to Alphabet, are projected to invest over $725 billion in hardware this year alone.
TSMC is not alone in seeking proximity to key clients. Micron Technology, an Idaho memory designer rivaling Samsung and SK Hynix for Nvidia's business, is investing $250 billion to boost domestic capacity.
Intel, based in Santa Clara near Nvidia, is also striving to attract clients seeking advanced manufacturing of AI accelerators.
Separately, Elon Musk, head of Tesla and SpaceX, is launching a chipmaking venture of his own, pointing to TSMC's inability to satisfy all AI-driven demand for the foreseeable future.
TSMC raised its 2025 capital expenditure target to a minimum of $60 billion.
After TSMC's results, technology stocks worldwide declined on ongoing worries about whether massive AI investments will yield significant returns.
Huang said on Friday that despite U.S. expansion, TSMC will continue to deploy its most advanced technology in Taiwan first, because creating next-gen chips demands tight coordination between R&D and manufacturing. Huang also reaffirmed TSMC's earlier declaration that it will not yet adopt ASML's most advanced extreme ultraviolet lithography machines, each costing over €350 million ($400 million), citing expense as a factor.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets
