TSMC has poured $165 billion into Arizona since 2020. Now the Taiwan-based foundry is exploring a Texas fab — its first American expansion outside Phoenix.

The Hsinchu-based company confirmed the Texas talks to SFGATE, offering no details on site, capacity, or process node. That original Arizona commitment — a $12 billion 5nm fab announced in May 2020 — has since expanded into multiple plants and an R&D center, anchored by $6.6 billion in direct CHIPS Act funding plus up to $5 billion in loans.

Why Texas?

Geography leads the list. Arizona sits inland; Texas offers a Gulf Coast shipping corridor, Houston port logistics, and an existing semiconductor ecosystem anchored by Samsung's Austin foundry and its repeatedly delayed $17 billion Taylor fab, originally planned for 4nm production and now pushed into 2026 or later.

Labor follows. UT Austin and Texas State feed an engineering workforce Samsung already exploits. Power matters too: Texas's deregulated grid offers industrial rate structures Arizona cannot match.

The commercial case looks customer-led. Industry estimates put roughly 92% of leading-edge logic capacity in Taiwan today. Apple, NVIDIA, and AMD have all pressed TSMC for onshore supply since Congress passed the CHIPS and Science Act in August 2022, earmarking $52.7 billion for domestic semiconductor manufacturing. Early-2025 tariff threats on imported chips added commercial urgency.

What does this signal for capacity?

CEO C.C. Wei has framed TSMC's US strategy as customer-driven, not subsidy-driven. The pattern across TSMC's global expansion suggests Texas won't chase the 2nm-and-below frontier. TSMC's Dresden fab, budgeted around $4 billion for automotive and industrial nodes, targets late 2027. The second Kumamoto plant, built with Sony and KKR, follows the same template.

If Texas fits that template, expect specialty processes for automotive, industrial, and IoT customers — segments where Texas proximity could shave logistics costs and align with regional data center demand. Cutting-edge 2nm and A16 work stays in Taiwan and Arizona.

Mature-node expansion in Texas would not directly address the 92% Taiwan concentration at 7nm and below. But it would serve customers willing to pay a premium for US-sourced wafers in segments where Samsung, GlobalFoundries, and Intel 18A all compete.

What's next for the Texas talks?

The Texas exploration arrives at a sensitive moment for TSMC's board. The Arizona commitment has grown more than tenfold from its original $12 billion scope. Adding Texas adds capital intensity TSMC must justify to investors accustomed to Taiwan-only margins.

The Texas announcement, when it comes, will test whether the customer-first expansion strategy survives contact with US subsidy politics. State governors from Ohio to New York have courted TSMC since 2022; Texas now leads the field. The pattern looks familiar — TSMC plays states against each other while securing the most generous package. Arizona ultimately delivered $6.6 billion in direct funds plus tax credits. Texas will likely demand matching terms.

Globally, TSMC's geographic spread now covers Taiwan, Arizona, Dresden, Kumamoto, and potentially Texas — five jurisdictions for a company that built its first three decades on a single island. Each new fab brings different labor laws, environmental rules, and IP regimes.

Whether Texas moves from exploration to formal announcement will depend on state-level incentives, water rights — a perennial Southwest friction point — and customer offtake commitments. None of those details accompanied the SFGATE disclosure. Watch for board-level approvals in 2025, state subsidy packages, and anchor customer contracts. The bigger question: does TSMC's American bet extend to two anchors, or does Arizona remain the sole frontier?