Japanese gov't to invest over ¥1 tril in chipmaker Rapidus

The Japanese government will provide over ¥1 trillion in investment and subsidies to chip venture Rapidus between fiscal 2026–27 as part of a broader push to secure domestic semiconductor supply chains; the trade ministry will invest ¥100bn this fiscal year and ~¥150bn in fiscal 2026, plus subsidies of ~¥630bn (fiscal 2026) and ~¥300bn (fiscal 2027), on top of ¥1.7tn in pledged subsidies and ~¥100bn of private capital. Rapidus, backed by eight major Japanese firms, plans a total ¥7tn investment program, aims to mass-produce 2nm chips in H2 fiscal 2027 with 1.4nm and 1nm following, and targets an IPO around fiscal 2031, reflecting industrial-policy-driven support to meet rising AI-related chip demand.

Analysis

Winners will be semiconductor capital-equipment and materials providers that supply EUV/etch/wafer/test kit chains (higher order books, stronger pricing power through 2025–28); losers are less-capitalized foundries and commodity-node suppliers that face margin squeeze if advanced-node capacity is subsidized into the market. Market share shifts will favor suppliers already qualified for advanced-node manufacturing (ASML, Tokyo Electron, Applied/Lam, SUMCO, JSR) because qualification lead times are 12–36 months and switching costs are high, compressing pricing leverage for second-tier vendors.

Major tail risks are technical failure to hit 2nm yields, geopolitical export controls limiting access to high‑NA EUV, and cost overruns triggering additional state support or project retrenchment; any of these would create a >30% re-rating in related small-cap suppliers. Immediate market moves will be muted (days); order-book and supplier revenue rerates should materialize over 6–24 months; structural impacts and IPO/valuation resets play out over 3–7 years.

Tradeable implication: bias into high-quality equipment/material names with 12–36 month convexity (buy LEAPS or call spreads) and avoid or underweight speculative domestic fab suppliers lacking equipment/web of customers. Key catalysts to watch for trade entry/exit are ASML high‑NA shipment dates, Rapidus pilot yield announcements (H2 FY2027 target window), and Japan’s FY2026–27 budget tranches.

Consensus misses the operational difficulty and skilled‑labor bottleneck; the market may be overpaying domestic names without technical accreditation. Historical parallels (state-led capex in solar/LCD) show subsidy-driven capacity can produce price deflation and stranded assets — favor upstream equipment leaders over end‑fab hopefuls to avoid mispriced execution risk.

Key Decisions for Investors

• Establish a 2–3% portfolio long split (60/40) between Tokyo Electron (8035.T) and Applied Materials (AMAT) with a 12–24 month horizon; take profits at +30%, stop-loss at -12% to capture the capex cycle while limiting execution risk.
• Allocate 1–1.5% to a 24–36 month call-spread on SUMCO (3436.T) (buy 20% OTM LEAP calls and sell 40% OTM calls) to play wafer demand while capping premium; exit or roll if ASML high‑NA shipping is delayed beyond Q4 2026.
• Implement a pair trade: long Advantest (6857.T) 1.5% vs short SMIC (688981.SS) 1.5% (or equivalent via options if borrow constrained) over 12–36 months to express relative exposure to advanced-node test demand versus lower-quality foundry risk; rebalance on each Rapidus milestone or ASML license update.
• Buy tail protection: allocate 0.5–1% to 12–18 month puts on EWJ or purchase a cheap downside collar on Japan-equity exposure; increase protection if Rapidus misses H2 FY2027 production target or if ASML announces export restrictions affecting high‑NA shipments.