Japan's Semiconductor Revival Strategy and the Conditions for Rapidus's Success

When assessing Japan’s semiconductor industry, it is necessary to distinguish among its historical share of the global market, current advanced logic production capacity, competitiveness in materials and equipment, and government support. Japan clearly lost its status as a comprehensive semiconductor powerhouse of the 1980s, but interpreting this directly as the disappearance of the entire industry or South Korea’s permanent superiority overlooks important facts.

This report focuses on the competitiveness of Japan’s production bases and Rapidus’s 2-nanometer plan, separating verifiable facts from projections that have yet to be validated.

What Japan’s Semiconductor Industry Lost and What It Retained

Japanese companies accounted for a very large share of the global semiconductor market in the 1980s, but subsequently lost leadership in memory and advanced logic. This resulted not from a single cause but from a combination of factors.

• Their product and investment strategies were slow to respond as the market shifted toward PCs and mobile devices.
• The vertically integrated general electronics company model failed to adapt properly to the division of labor between fabless companies specializing in design and foundries specializing in contract manufacturing.
• Business withdrawals and consolidation continued amid memory price competition and the burden of large-scale capital investment.
• Japanese companies also fell behind U.S., South Korean, and Taiwanese companies in decision-making speed, business restructuring, software ecosystems, and the acquisition of global customers.

However, Japan’s semiconductor industry has not disappeared entirely. Japanese companies retain important positions across multiple areas of the supply chain, including image sensors, silicon wafers, photoresists, cleaning, inspection, and deposition equipment, and precision components. Japan’s weakness therefore lies primarily in advanced logic design and high-volume manufacturing capacity; it does not mean that the country lacks the entire industrial foundation encompassing materials and equipment.

Smartphone Sales and Digital Supremacy Are Not the Same Metric

Samsung’s decision to foreground the Galaxy brand in Japan for a period and then resume using the SAMSUNG name on some new products beginning in 2023 can be viewed as a change in brand strategy. Instances in which Samsung ranked among the top Android sellers during certain periods or through certain sales channels also indicate that Japanese consumers have become more receptive to South Korean brands.

However, this should not be expanded into a claim that Samsung ranks first in Japan’s overall smartphone market or that digital supremacy between South Korea and Japan has definitively reversed. Smartphone statistics differ according to how they are measured.

• Shipments are the number of units supplied by manufacturers to distribution channels.
• Sales are the number of devices actually sold to consumers and vary depending on the scope of the channels surveyed.
• Web usage share is not sales volume, but the share of internet traffic generated by devices currently in use.
• Rankings within Android and rankings for the overall market including the iPhone are different metrics.

Changes in perceptions among younger generations may have contributed to higher sales, but it is difficult to identify generational turnover as the sole cause without research on purchasing motivations by age group. Prices, mobile carrier promotions, foldable products, exchange rates, launch timing, and distribution inventory also play a role.

“Digital supremacy” is likewise not a single officially recognized statistic. South Korea is strong in memory, smartphones, displays, and some digital services, while Japan is strong in materials, equipment, sensors, and industrial technology. Taiwan dominates advanced foundries, while the United States has overwhelming influence in semiconductor design, software, and cloud services. Competitiveness among countries must be assessed using separate metrics for each area.

Why South Korean Companies Do Not Build Large Fabs in Japan

The fact that South Korean companies have not announced large-scale advanced production plants in Japan is not sufficient to conclude that they are completely turning away from the country. Semiconductor companies use different criteria to locate production fabs, research and development facilities, packaging operations, and customer support bases. Research and development or collaborative investment in Japan is a separate choice from investment in a large wafer fab.

Domestic Demand

An advanced fab does not necessarily have to consume all of its output locally. Semiconductors are exported to global markets, and foundries operate based on customer orders. The explanation that “a fab cannot be built because there is insufficient demand for 2-nanometer chips in Japan” is therefore incomplete.

However, having major customers nearby is advantageous for joint development, quality response, and demand forecasting. Japan has demand in the automotive, sensor, and industrial equipment sectors, but its domestic fabless base placing large-volume orders for cutting-edge processes, such as smartphone application processors or AI accelerators, is relatively weak. This is why Rapidus must secure overseas customers.

Logistics and Industrial Clusters

Finished chips are small and valuable, making transportation costs less burdensome than for automobiles or steel. However, this does not mean that logistics costs are irrelevant to fab location decisions. Wafers and materials require contamination, shock, and delivery-time management, while access to equipment maintenance personnel, chemicals, components, packaging facilities, and customer engineering teams is also important.

Semiconductor locations are determined by a combination of the following factors rather than transportation costs alone.

1. Clusters of skilled engineers and partner companies
2. Stable supplies of electricity and ultrapure water
3. Equipment and materials supply chains and maintenance speed
4. Government subsidies and taxes, and the time required for permits and approvals
5. Customer accessibility, intellectual property protection, and geopolitical risk

Earthquake Risk

Japan’s earthquake risk is a genuine operational risk. Microvibrations, power outages, damage to water pipelines, and logistics disruptions can cause production setbacks and wafer losses. They also increase insurance premiums and the costs of earthquake-resistant design, inventories, and supply-chain redundancy.

Even so, it cannot be concluded that all of Japan is unsuitable for semiconductor production. Taiwan also faces significant earthquake risk but operates the world’s largest foundry production base. Modern fabs use seismic isolation structures, emergency power, and automated equipment shutdown and restart procedures. The key issue is not whether risk exists, but the level of risk at a particular site and the cost required to reduce it.

Power Supply

Advanced fabs continuously consume enormous amounts of electricity, making prices, carbon emissions, grid capacity, and outage quality important. Since the Fukushima accident, Japan has experienced electricity cost and supply pressures due to reduced nuclear power generation and reliance on imported fuels, and it also faces limited interconnection among regional grids.

However, characterizing Japan as a country with chronic power shortages or saying that operating large fabs there is effectively impossible would also be an exaggeration. Some nuclear power plants have restarted, and power supply conditions vary by region and season. Actual investment decisions should examine not the total amount of generating capacity, but the grid connection capacity available in the relevant region, long-term power contracts, backup power, renewable energy procurement, and plans to expand the transmission network.

What TSMC’s Kumamoto Investment Demonstrates

JASM, established by TSMC with Sony Semiconductor Solutions and others, is expanding production in Japan through its Kumamoto plant. Publicly announced plans include multiple processes intended to meet demand from the automotive, industrial, consumer, and high-performance computing sectors.

This case demonstrates two things simultaneously.

• Japan can attract global foundries by combining subsidies, existing customers, materials and equipment supply chains, and a regional industrial base.
• The Kumamoto investment does not itself mean that Japan has recovered independent, cutting-edge 2-nanometer capabilities. JASM is an overseas production base that uses TSMC’s technology and operating systems.

There is also no reason why every chip produced by TSMC at its Japanese plant must be consumed within Japan. Production bases are part of global supply chains, and their economic viability is assessed based not only on local domestic demand but also on exports and customer portfolios.

Structure of Rapidus’s 2-Nanometer Plan

Rapidus was established in 2022 with investments from eight Japanese companies, including Toyota, Sony Group, NTT, SoftBank, Kioxia, NEC, Denso, and MUFG Bank. It is building a production base in Chitose, Hokkaido, and aims to begin high-volume production of 2-nanometer-class logic semiconductors in 2027.

Rather than sequentially catching up from older generations, Rapidus’s strategy is to enter the advanced generation directly by leveraging cooperation on IBM’s 2-nanometer-class technology and an international research network. This may shorten development time, but high-volume manufacturing experience is not transferred automatically.

In the schedule disclosed through 2025, operation of the pilot line and prototype validation were presented as important milestones. However, producing transistors or test wafers, providing customer samples, achieving certified high-volume manufacturing, and conducting profitable mass production are all separate stages.

Five Conditions That Will Determine Rapidus’s Success

1. Process Technology Reproducibility

Performance is not determined by the 2-nanometer label alone. Transistor density, speed, power consumption, design rules, and process stability must be evaluated together. Even if Rapidus receives underlying technology from IBM, it must repeatedly reproduce the same results using the equipment and material conditions in Chitose.

2. Economically Viable Yield

Yield indicates how many saleable chips are produced from the wafers processed. Reducing wafer-to-wafer variation and lowering defect density is much more difficult than successfully producing a single test chip. Low yields increase the cost per chip, leading customers to choose proven competing foundries.

3. Design Ecosystem

Foundry customers do not purchase manufacturing processes alone. They need process design kits, verified semiconductor design assets, electronic design automation tools, packaging, testing, and technical support. For Rapidus to turn customer designs into actual chips, it must build a mature ecosystem with global software, IP, and packaging companies.

4. Key Customers and Utilization Rates

Advanced fabs have extremely high fixed costs and therefore require stable orders and high utilization rates. Because Rapidus is unlikely to fill its plant with Japanese customers alone, it must secure overseas customers that design chips for AI, data centers, automobiles, and telecommunications.

There is insufficient evidence to conclude that it has “no customers at all,” but securing key customers that guarantee long-term, high-volume purchases and obtaining sufficient order volumes are separate issues. Memoranda of understanding on technology cooperation or joint prototype development are not equivalent to confirmed high-volume production orders.

5. Continuity of Funding and Talent

Advanced processes require large-scale research and development and equipment investments every year, even after plant construction. Government support can reduce initial risk, but private investment and revenue must continue for the business to become self-sustaining. Rapidus must not only recruit people with overseas experience but also combine personnel in process integration, equipment, yield, and customer support into a unified operating organization.

Metrics for Assessing Rapidus’s Chances of Success

Rather than assigning a definitive numerical probability of success, it is more accurate to monitor publicly disclosed milestones continuously.

A Realistic Assessment of Japan’s Semiconductor Policy

Rapidus is neither a project proven to be technologically impossible nor one guaranteed to succeed through government funding alone. If successful, Japan could gain a new option in the advanced logic supply chain and strengthen ties with its materials and equipment industries. If it fails, large amounts of public funding and production facilities could remain without a sufficient customer base.

The policy’s success or failure should not be assessed simply by whether Rapidus “made a 2-nanometer chip” in 2027, but by the following questions.

• Did it meet the performance and reliability requirements of customers?
• Can it repeatedly manufacture chips at a competitive price?
• Can it invest in subsequent processes even if government support declines?
• Are Japan’s workforce and design ecosystem continuing to expand?
• Does supply-chain diversification provide strategic value greater than the resulting increase in costs?

Conclusion

Japan faces a large gap in high-volume advanced logic manufacturing, but it possesses foundations in materials, equipment, sensors, manufacturing customers, and capital. TSMC’s Kumamoto plant is an example of combining these strengths with overseas technology, while Rapidus is a much riskier attempt to rebuild independent advanced foundry capabilities.

Japan’s limited domestic demand, earthquakes, and power issues are all risks that must be examined, but none automatically determines failure. Conversely, government subsidies and technology partnerships alone do not create yield, customer trust, or high-volume manufacturing expertise.

Therefore, rather than concluding that “the revival of Japan’s semiconductor industry is impossible” or that “the reversal of digital supremacy between South Korea and Japan is irreversible,” it is more accurate to assess Rapidus’s progress using measurable indicators such as process performance, yield, customer orders, utilization rates, and private funding.