A Sixth Approval, and a Widening Gap
On July 2, 2026, Taiwan's Ministry of Economic Affairs (MOEA) cleared a US$20 billion capital injection by TSMC into TSMC Arizona, its wholly owned U.S. subsidiary, to fund a 12-inch wafer fab and an advanced packaging plant. The Department of Investment Review approved it alongside eight other projects, including outbound investments from Nanya Technology, Quanta Computer, and Lite-On Technology, at a single Thursday meeting (Focus Taiwan). It was the sixth time Taipei's investment screen has cleared a TSMC U.S. capital transfer, pushing the cumulative approved total to US$44 billion.
That figure sounds substantial until it is set against what TSMC has actually promised Washington. On July 16, 2026, the company confirmed an additional $100 billion U.S. investment, following a January 2026 U.S.-Taiwan trade and investment agreement, bringing its total planned American commitment to $265 billion across twelve leading-edge fabrication and packaging facilities (NIST/Commerce Department). Do the arithmetic and Taiwan's Investment Commission has formally cleared roughly 17 percent of the capital TSMC has told the U.S. government it intends to spend. The remaining 83 percent exists only as a public pledge, still awaiting the tranche-by-tranche approvals that Taiwanese law requires before a single additional dollar can legally leave the country.
The Legal Machinery Behind the Rubber Stamp
This review process is not a rubber stamp by design, even if it functions like one in practice. Taiwan amended Article 22 of its Statute for Industrial Innovation, promulgated May 7, 2025, to give MOEA authority to condition or reject outbound investment approval not just on dollar thresholds but on "the nature, destination and strategic importance of the investment," explicitly including national security and defense considerations (Industrial Development Administration, MOEA). A companion Article 67-3 backs this with real teeth: initial fines of NT$50,000 to NT$1 million for unauthorized outbound capital transfers, escalating to NT$500,000-10 million per violation for continued noncompliance, plus mandatory divestment orders.
This is a serious statutory framework, built in the shadow of concerns about capital and technology flight to China, then repurposed to also govern the far larger and more politically fraught question of how fast Taiwan's most important company should be allowed to relocate manufacturing capacity to the United States.
Steelmanning the Screen
The case for keeping this review in place is genuinely strong, and skeptics of industrial policy should not wave it away. TSMC is not an ordinary company to Taiwan: it manufactures roughly 90 percent of the world's most advanced logic chips, and its concentration on the island functions as a deterrent against Chinese aggression — the so-called "silicon shield" thesis that argues China, and the world, have too much to lose from a war that disrupts Taiwanese fabs to risk one. A government that let its crown-jewel manufacturer offshore capital and capacity unchecked, at whatever pace Washington or Beijing pressure dictated, would be abdicating a legitimate economic-security function. Public, criteria-based review — rather than backroom deal-making — is also more transparent than the alternative, and Taiwan has real precedent for corporate capital flight distorting its industrial base during earlier waves of cross-strait investment.
Why the Gap Undercuts the Chokepoint Story
But the 17 percent figure tells a more mundane story than "national security screen." TSMC's $265 billion figure was not a request pending Taiwanese permission — it was announced by the company and the U.S. Commerce Department as a done political deal, negotiated at the head-of-government level in January 2026, months before Taipei's Investment Commission had cleared even a third of the prior $165 billion tranche. The review process is not shaping whether or how much capital moves; that decision is made in Washington and TSMC's boardroom. What MOEA controls is sequencing and paperwork — useful for monitoring, largely irrelevant to outcome.
That is, on balance, the right design. A screening regime that actually tried to block or meaningfully slow $265 billion in already-announced investment would trigger a fight with Washington that Taiwan cannot afford and would not win, given how central the TSMC relationship now is to U.S.-Taiwan security cooperation. A regime that merely logs, monitors, and occasionally attaches conditions — while leaving the real decision to market and diplomatic forces — lets Taipei preserve formal sovereignty over its most strategic firm without picking a fight it cannot win.
The Risk Worth Watching
The proportionate-regulation case for this framework holds only as long as MOEA keeps functioning as a monitor rather than a lever. Article 22's national-security language is broad enough that a future government, under different cross-strait or U.S. political pressure, could use the same statute to slow-walk approvals as leverage in an unrelated dispute — turning a transparent administrative process into a geopolitical chokepoint overnight. Nothing in the current pattern of six swift, unremarkable approvals suggests that is happening. But the legal authority for it now exists, and it is worth Taiwanese lawmakers and TSMC's other host governments watching whether future approvals keep pace with this one, or start to lag for reasons that have nothing to do with the criteria on the statute books.