In the 1980s, Japan made over half the world's chips. Then came decades of decline. Now the government has set a target that would have been unthinkable five years ago: grow domestic semiconductor sales from ¥5 trillion ($33 billion) in 2020 to ¥40 trillion ($260 billion) by 2040. Behind it sit a state-backed 2nm startup in Hokkaido, TSMC's $17 billion 3nm fab in Kumamoto, and a $13 billion display plant pitched to Washington. Here's the full picture.

The $260 Billion Roadmap

On March 10, 2026, Japan's Growth Strategy Council, chaired by Prime Minister Sanae Takaichi, met at the prime minister's office and designated 61 products and technologies for priority public and private investment. For 27 of them, including AI robotics and the semiconductors that make it work, it published draft roadmaps. The headline number for chips: ¥40 trillion ($260 billion) in domestic semiconductor sales by 2040.

That sits on top of an existing target of ¥15 trillion ($100 billion) by 2030. Japan's domestic chip output was worth roughly ¥5 trillion ($33 billion) in 2020 and stands at about ¥8 trillion today, so ¥40 trillion means eight times the 2020 figure and five times the current one. The plan will be formally incorporated into the national growth strategy this summer, with the government setting direction to encourage private-sector investment.

Driving this ambition is the explosive growth in AI and data center demand. In its autumn forecast published in December 2025, the World Semiconductor Trade Statistics (WSTS) put the 2026 global chip market at $975.5 billion, up 26.3% from 2025, with Japan at roughly $50.1 billion. WSTS has since revised that 2026 figure sharply upward, to $1.5112 trillion, in its spring forecast of June 2026. The demand estimates keep moving in one direction.

To capture this growth, the government is accelerating the development of cutting-edge semiconductor R&D hubs, aiming for an integrated domestic ecosystem spanning AI chip design, manufacturing, and advanced packaging.

Rapidus: Japan's 2nm Moonshot in Hokkaido

The flagship of Japan's semiconductor strategy is Rapidus, a state-backed foundry established in 2022. Its mission: mass-produce the world's most advanced 2nm chips at the IIM-1 facility in Chitose, Hokkaido, starting in the second half of fiscal 2027.

The milestones have come fast. The pilot line began operations in April 2025. By July 2025, Rapidus unveiled prototype 2nm GAA (Gate-All-Around) transistors that passed initial functionality tests, a critical technical achievement. Production capacity will start at 6,000 wafers per month, scaling to 25,000–30,000 wafers within the first year of operation.

On February 27, 2026, Rapidus closed ¥267.6 billion ($1.7 billion) in funding: ¥100 billion from the government's Information-technology Promotion Agency (IPA) and ¥167.6 billion from 32 private companies including Canon, Fujitsu, NTT, SoftBank, Sony Group, and Toyota. The private portion comfortably beat the ¥130 billion the government had penciled in. The IPA stake made the state Rapidus's largest shareholder with 11.5% of voting rights, plus a golden share carrying veto power over key decisions. In November 2025 the industry ministry said it would add roughly ¥1 trillion in support across fiscal 2026 and 2027, bringing cumulative state support to ¥2.9 trillion by the end of fiscal 2027. A broader framework pledges over ¥10 trillion in public support for AI and semiconductors across seven years.

What makes Rapidus unusual is its short turnaround manufacturing model. While conventional foundries take about 120 days to process a wafer, Rapidus aims for roughly 50 days, positioning itself as the fast-turnaround partner for AI startups and custom chip designers who prioritize speed over volume. A move to 1.4nm is pencilled in for around 2029.

TSMC in Kumamoto: A $17 Billion 3nm Surprise

The second pillar of Japan's chip revival is TSMC's expanding presence in Kumamoto, southern Japan.

TSMC's subsidiary JASM (a joint venture with Sony, Denso, and Toyota) began mass production of 12–28nm chips at its first Kumamoto fab in late 2024. But the real game-changer came on February 5, 2026, when TSMC CEO C.C. Wei personally visited Prime Minister Takaichi to announce a dramatic upgrade: the second Kumamoto fab would shift from 6–7nm to cutting-edge 3nm production.

The total investment jumped from roughly $12.2 billion to approximately $17 billion (¥2.6 trillion), making Kumamoto only the third location worldwide, after Taiwan and Arizona, where TSMC will produce 3nm chips. These are the same advanced semiconductors used in AI accelerators, flagship smartphones, and autonomous driving systems. Production is likely to begin around 2028.

The economic ripple effects in Kumamoto have been transformative. Over 170 companies have invested in or expanded operations in the region, with cumulative economic impact estimated at over ¥11 trillion ($72 billion) from 2022 to 2031. The Japanese government has already approved roughly ¥1.2 trillion ($8 billion) in subsidies for JASM, up to ¥476 billion for the first fab and up to ¥732 billion for the second, with additional support for the 3nm upgrade under consideration.

Importantly, the government views TSMC's 3nm and Rapidus's 2nm as complementary rather than competing initiatives. TSMC offers proven high-volume manufacturing for mainstream AI and mobile applications, while Rapidus targets fast-turnaround, small-batch production for cutting-edge custom chips.

JDI's $13 Billion US Display Factory

In a twist that connects semiconductor strategy with US-Japan trade dynamics, the Japanese government has approached Japan Display Inc. (JDI) about operating a cutting-edge display factory in the United States, with an estimated project scale of $13 billion (¥2 trillion).

The proposal emerged as a new candidate for Japan's $550 billion investment and loan package to the US, a commitment made under the Takaichi-Trump framework agreement. JDI had already announced a strategic partnership with OLEDWorks to build an advanced OLED manufacturing hub in the US for defense, automotive, and medical display applications.

While JDI is traditionally a display company, its "BEYOND DISPLAY" strategy includes a push into advanced semiconductor packaging and 3D chip integration technology, making it a potential bridge between Japan's display expertise and the broader semiconductor ecosystem.

How Japan Compares: The Global Chip Subsidy Race

Japan's strategy doesn't exist in a vacuum. Every major economy is racing to secure domestic semiconductor capabilities.

United States (CHIPS Act, 2022): $39 billion in manufacturing subsidies over five years, plus $11 billion for R&D, and a 25% investment tax credit. Major recipients include Intel, TSMC's Arizona complex ($165 billion total commitment), and Samsung.

European Union (EU Chips Act, 2023): A target of €43 billion (about $50 billion) in combined public-private investment by 2030, aiming to raise the EU's share of global chip production to 20%.

South Korea: Enhanced tax credits for semiconductor investment (20% for large firms, 30% for SMEs), ₩17 trillion ($13 billion) in low-interest loans, and extended R&D tax incentives through 2031. Samsung and SK Hynix are the primary beneficiaries.

Taiwan: Home to TSMC's core operations, Taiwan maintains approximately 85% of the world's sub-5nm manufacturing capacity. The government provides generous R&D tax credits and university partnerships.

China: Over ¥10 trillion ($65 billion) invested through national and regional semiconductor funds, plus a new ¥7 trillion ($45 billion) fund established in 2024. Despite US-led export restrictions on advanced equipment, China is aggressively pursuing self-sufficiency.

Japan's distinguishing feature is the combination of speed and breadth. The country invested approximately ¥4 trillion ($26 billion) in semiconductor support over just three years (2021–2024), the largest among advanced economies relative to GDP. Moreover, Japan dominates semiconductor manufacturing equipment (Tokyo Electron, Advantest, Screen Holdings) and materials (Shin-Etsu Chemical, SUMCO, JSR), giving it irreplaceable leverage in the global supply chain.

The Three-Layer Strategy

Japan's semiconductor approach operates on three distinct layers:

Layer 1: Strengthening existing manufacturing: Partnerships with TSMC (Kumamoto), Kioxia (Yokkaichi), and Micron (Hiroshima) secure reliable production of mature 12–28nm chips for automotive, industrial, and consumer applications.

Layer 2: Establishing next-generation capabilities: Rapidus (2nm) and TSMC Kumamoto Fab 2 (3nm) bring cutting-edge logic chip production to Japanese soil, supported by US-Japan technology partnerships with IBM and imec.

Layer 3: Investing in future technologies: NTT's photonics-electronics convergence (IOWN), diamond semiconductors, and quantum computing chips represent Japan's bets on post-2030 breakthroughs.

Challenges Ahead

The ¥40 trillion target faces serious headwinds.

Talent shortage is the most pressing concern. Demand for semiconductor engineers is projected to hit all-time highs by 2027, requiring major university curriculum reforms and aggressive international recruitment.

Yield improvement at Rapidus remains an open question. Manufacturing 2nm GAA transistors is extraordinarily difficult, and even TSMC and Samsung have struggled with it. Whether Rapidus can achieve commercially viable yields by its fiscal 2027 mass production target is uncertain.

Return on public investment will face scrutiny. With ¥2.9 trillion of state support headed to Rapidus by the end of fiscal 2027 and up to ¥1.2 trillion for JASM, taxpayers will expect measurable economic returns.

Infrastructure strain in Kumamoto and Chitose, both relatively small cities, is already evident in housing shortages, traffic congestion, and competition for water and electricity resources.

Can Japan Reclaim Its Chip Crown?

Japan once commanded over 50% of the global semiconductor market. Trade friction with the US, a failure to adopt the fabless model, and insufficient investment in leading-edge logic manufacturing led to decades of decline.

But the confluence of geopolitical risk, AI-driven demand, and enormous state commitment has created a once-in-a-generation opening. The ¥40 trillion target is ambitious, and it is backed by concrete projects, real money, and international partnerships. The question is no longer whether Japan is serious about semiconductors. It is whether the execution can match the ambition.

What's the semiconductor landscape like in your country? Is your government investing in chip manufacturing? We'd love to hear your perspective, tell us about the chip industry where you live!

References