Japan METI JASM second-fab subsidy: 732 billion yen for TSMC Kumamoto Fab 2

Mechanism

The 732 billion yen subsidy is funded under the Amended Act on Facilitation of Specified Semiconductor Production Facilities (the "5G Promotion Act" framework that already underwrites Fab 1, Rapidus IIM-1 in Chitose, and Micron Hiroshima HBM3e). METI selects designated projects via certification of a "system development plan"; once certified, the operator draws subsidies against actual capital expenditures up to a notified ceiling. Fab 2's certified ceiling is 732 billion yen. Together with the 476 billion yen Fab 1 ceiling, the combined Kumamoto site is supported by up to ~1.208 trillion yen of public capex co-financing, approximately 47-50% of total project cost (combined investment exceeds US$20 billion).

JASM's shareholder structure for the combined Fab 1 + Fab 2 entity is TSMC ~86.5%, Sony Semiconductor Solutions 6.0%, Denso 5.5%, and Toyota Motor 2.0% — Toyota's stake was added in February 2024 alongside the Fab 2 expansion announcement. Fab 2 escalates the technology mix: Fab 1 produces 12-28nm mature-node logic; Fab 2 adds 6/7nm advanced logic plus 40nm legacy nodes, with combined site capacity exceeding 100,000 12-inch wafers per month. End-markets are automotive, industrial, consumer, and HPC.

METI imposed two structural conditions explicitly tied to Fab 2 at the 24 February announcement:

1. Production durability: Fab 2 must remain in production for at least 10 years after mass-production start. This locks in long-dated upstream tooling and materials demand and prevents an early write-off if economics deteriorate. 2. Domestic procurement: at least 50% of silicon wafers and at least 50% of components/materials must be procured from Japan-based suppliers. This converts the subsidy into demand-pull for Shin-Etsu Chemical, SUMCO, JSR, Tokyo Ohka Kogyo, Tokyo Electron, Screen Holdings, and the broader Kumamoto cluster (~90 facilities, ~10,700 jobs).

Severity is set at 4 (same as Fab 1): the largest single direct manufacturing subsidy in Japanese history, decisive for advanced-node geographic diversification of TSMC's leading-edge capacity outside Taiwan and meaningful for a half-dozen Japanese specialty-materials suppliers.

Downstream implications

• EWJ (Japan broad-market): Positive structural signal. Sony,

Denso, Toyota direct stakes; Shin-Etsu, SUMCO, Tokyo Electron, Screen Holdings, JSR, Tokyo Ohka Kogyo upstream demand-pull from the >=50% domestic-procurement clause. Yen-weakness risk asymmetric here — capex denominated in yen, output exported in USD.

• SOXX / SMH (semiconductor majors): Positive for TSMC

diversification premium — first overseas advanced-node (6/7nm) capacity outside Taiwan/Arizona. Reinforces the trilateral Taiwan-Japan-US foundry footprint vs. China-localised capacity.

• EWT (Taiwan): Marginally negative on a long-horizon basis —

Fab 2 nibbles at Taiwan's monopoly on TSMC leading-edge capacity and accelerates the geopolitical hedge logic. Magnitude is small vs. TSMC's Taiwan capex baseline.

Open questions

• Will METI further upsize the framework if TSMC commits to a third

Kumamoto fab (3nm)? The December 2024 deferral of the third-fab decision suggests the answer is conditional on Trump-era trade policy and TSMC Arizona ramp pacing.

• How rapidly does the 50%-domestic procurement floor bind in

practice? Shin-Etsu/SUMCO can supply silicon, but advanced photoresists and CMP slurries remain Japan-strong; EUV-grade reticles and certain etch chemistries are weaker links.

• Does the Fab 2 6/7nm node spec stick, or does it migrate to 3nm

as TSMC's roadmap moves down? The April 2026 Taiwan-government approval for a 3nm upgrade at the second Japan fab suggests the node may actually advance ahead of plan.