Chips in Japan: Industrial policy, decline and renewal

Japan, once the global leader in semiconductor manufacturing, has re-entered the centre of global industrial policy debates as it attempts to rebuild domestic chip capabilities in an era of escalating geopolitical rivalry, supply-chain fragility and rapid technological change. This paper examines Japan’s semiconductor trajectory over the past half-century: from global dominance in the 1980s, to steep decline in the 1990s and 2000s, to its current large-scale industrial policy interventions.

Japan’s semiconductor leadership in the 1970s and 1980s emerged from manufacturing excellence, strong consumer electronics demand, long-term financing via keiretsu structures and coordinated R&D initiatives such as the VLSI project. Japan dominated memory chips, particularly DRAMs, and overtook the United States by the mid-1980s. However, structural and institutional rigidities undermined Japan’s position as the industry shifted toward the fabless–foundry model, shorter product cycles and logic-chip-centred architectures. The 1986 US–Japan Semiconductor Agreement, yen appreciation after the Plaza Accord and the rapid rise of Korean and Taiwanese competitors further weakened Japanese firms. By the early 2000s, Japan had lost its technological edge and concentrated largely on mature-node production, with its share of global semiconductor revenues falling below 10 per cent.

Japan’s long-term innovation constraints included over-coordination, excessive standardisation, weak start-up formation, low labour mobility and the late adoption of university–industry research platforms – factors that hindered the transition to new architectures and business models. The ageing engineering workforce and outflow of talent to Taiwan, Korea and China further eroded Japan’s ecosystem. Although Japanese firms remained globally significant in equipment and materials, they struggled to maintain leadership in high-end manufacturing and design.

The COVID-19 pandemic, supply-chain shortages in automotive and industrial chips, escalating US–China strategic rivalry and heightened concern over the weaponisation of interdependence have collectively prompted Japan to adopt a renewed and expansive semiconductor strategy. Between 2021 and 2023, Japan committed nearly JPY 3.9 trillion (US$27 billion) – proportionally the largest semiconductor investment package among advanced economies – to rebuild domestic supply-chain resilience, reduce exposure to geopolitical shocks and restore competitiveness in advanced manufacturing.

Japan’s strategy reflects two complementary logics: (i) securing stable supply of legacy and mid-range logic chips critical to automotive and industrial sectors; and (ii) attempting to re-enter the global technological frontier in advanced logic chips for AI and high-performance computing. The approach represents a departure from earlier inward-looking “all-Japan” coordination: it now embraces foreign firms, global technology partnerships, and integration into international production networks.

The centrepiece of Japan’s pragmatic, lower-risk strategy is TSMC’s Japan Advanced Semiconductor Manufacturing (JASM) fabs in Kumamoto, in which Sony, Denso and Toyota jointly invested. Supported by over US$8 billion in subsidies, the JASM fabs rapidly reached construction milestones, catalysing tens of billions of dollars in related investments and reinvigorating Kyushu as a semiconductor cluster. JASM aims to enhance supply security for automotive and industrial applications while embedding Japan within TSMC’s global ecosystem. It illustrates a model of market-aligned, foreign-partner-led industrial policy.

In contrast, the Rapidus project in Hokkaido embodies Japan’s high-risk frontier ambition: skipping several generations of technological progression to produce 2-nanometre logic chips by 2027. Backed by a consortium of eight major Japanese firms and supported by over US$11 billion in public funding, Rapidus relies heavily on technology transfer from IBM and IMEC. Its business model targets rapid-turnaround production for specialised AI and high-performance computing applications. However, the project faces major challenges: unprecedented technological leaps, limited private capital commitments, acute engineering shortages, high utility costs and the difficulty of attracting anchor customers to a new, unproven foundry.

Beyond these flagship projects, Japan is also supporting memory production (Kioxia/Western Digital), DRAM (Micron), back-end and packaging R&D, regional training consortia, the Leading-Edge Semiconductor Technology Centre, and expanding tax incentives for R&D, capital investment and IP commercialisation.

Japan’s semiconductor supplier networks are expanding; local universities are adjusting curricula and broader regional ecosystems are deepening. Yet for both Rapidus and JASM, the technological and commercial risks remain substantial. The leap to 2nm without an existing advanced-node base is unmatched globally. Workforce shortages, limited private co-investment, and the absence of committed high-volume customers raise concerns about viability. And even in the lower-risk project, JASM, Japan must address the global competition for legacy chips that dampens levels of investment and consumer demand.

At the global level, Japan’s strategy unfolds amid intensifying subsidy competition, raising risks of inefficient expenditure and fragmented production. Japan’s experience also underscores broader tensions in industrial policy at the technological frontier, where outcomes are highly uncertain and spillovers unevenly distributed.

The paper identifies several lessons for effective industrial policy:

1. 1. Openness and global partnerships are essential. Foreign-led and collaborative models foster competition, reduce insularity and accelerate technology transfer – addressing past weaknesses.
2. 2. Diversification across business models reduces systemic risk: pairing lower-risk foreign-led fabs with a high-risk domestic frontier project effectively hedges government semiconductor investments.
3. 3. Performance-linked funding is critical: milestones, matched private investment and transparent reporting could strengthen private accountability and reduce moral hazard.
4. 4. Ecosystem-oriented strategies outperform isolated flagship projects. But competitive ecosystems require reforms in talent development, labour mobility, start-up formation, university–industry collaboration and immigration policy.
5. 5. International cooperation on subsidies and technology governance can reduce costly subsidy races and support collective resilience.
6. 6. Independent reviews of subsidy-linked performance and technical – not political – exit criteria from government support are crucial for evaluating the ongoing feasibility of semiconductor investments.

Japan’s semiconductor revival is one of the most ambitious industrial policy undertakings among advanced economies. Its success will depend on sustaining openness, fostering private-sector leadership, strengthening human capital, maintaining competitive neutrality and ensuring that industrial interventions remain market-aligned. While JASM provides a promising template for effective, internationally embedded industrial strategy, Rapidus represents a bold but uncertain attempt to restore Japan to the technological frontier. Together, these dual pathways offer substantive insights into how advanced economies can design industrial policy that balances resilience, competitiveness and global integration in an era of geoeconomic fragmentation.