Japan's Double Bet on Semiconductor Revival: The Kumamoto Model and the Rapidus Path

Japan's Semiconductor Revival Double Bet: The Kumamoto Model and the Rapidus Path

From the Lost Decade to Economic Security

In the 1980s, Japan's semiconductor industry stood at the top of the world, with giants like NEC and Hitachi nearly unrivaled in DRAM. But by the 1990s, the market shifted to logic chips, the rise of American technology, and cost pressures from yen appreciation left Japanese manufacturers lagging step by step, eventually withdrawing completely from the advanced process race after the 22/20nm node. Many economic historians regard this as the watershed of Japan's semiconductor lost era.

Then, under the impact of the global chip shortage in 2020, the Japanese government realized anew that semiconductors could not be entirely dependent on imports. In 2021, a strategy led by then-Prime Minister Fumio Kishida placed semiconductors at the core of economic security. The Act on the Promotion of Economic Security, passed in 2022, defined four pillars: ensuring stable supply of critical materials, protecting infrastructure, supporting advanced technology development, and establishing a patent secrecy system. It was within this policy framework that two high-profile projects came into being: TSMC's Kumamoto fab and Rapidus.

On the surface, both projects are "redeeming" Japan's semiconductors, but their strategic logic is almost diametrically opposed.

The Kumamoto Model: Attracting a Global Leader to Rebuild the Supply Chain

TSMC's Kumamoto fab is a bold attempt by the Japanese government to attract direct investment from the world's largest wafer foundry. The fab, with participation from local companies such as Sony and Denso, produces mature-process chips needed for automobiles, industrial equipment, and consumer electronics. Kumamoto is not pursuing the most cutting-edge nanometer nodes; rather, it aims to fill Japan's domestic supply gap for mature chips and bring chip manufacturing back home, creating synergies with Japan's strong materials, equipment, automotive, and electronics industries.

Economist Jun Akaba pointed out at a Stanford APARC seminar that the value of the Kumamoto project lies not only in one factory, but in the chain reaction it triggers within the existing supply chain. Local equipment and materials suppliers are reincorporated into the global production network, and Japan's domestic technical workers and engineering talent gain a new training stage. This strategy of activating the internal system through foreign investment appears pragmatic and quick to deliver results.

However, criticism is equally sharp. Some question: why is the Japanese government spending huge subsidies to support a foreign company? If TSMC adjusts its investment strategy in Japan in the future due to political pressure or commercial interests, will Japan still end up with nothing? More fundamentally, the Kumamoto model has not reshaped Japan's own logic chip design capability; it has only made Japan a "high-quality capacity base" in the global foundry landscape.

Rapidus: An All-or-Nothing Expedition into Advanced ProcessesIn stark contrast to Kumamoto's pragmatism is Rapidus. Established in 2022, this new company is jointly funded by eight Japanese corporate giants including Toyota, Sony, and SoftBank, with the government making a rare move to directly "step in" with financial support. Its goal is exceedingly ambitious: allying with IBM and Belgium's imec (Interuniversity Microelectronics Centre), it aims to mass-produce 2-nanometer logic chips and bring Japan back to the table of global advanced semiconductors.

This is by no means an easy path. According to Jun Akaba's analysis, Rapidus remains at the 14-nanometer technology level, a huge gap from 2-nanometer mass production; it has not even yet acquired the key equipment required to make 2-nanometer chips. More critically, Rapidus must contend with TSMC, Samsung, and other giants for a slice of the key equipment procured from ASML, making the limited supply of advanced manufacturing equipment a potential bottleneck.

Supporters of Rapidus believe that only by taking on the most advanced process nodes can Japan cast off its fate as a "semiconductor supporting role." But detractors see this as a high-risk national gamble. They argue that even if Rapidus succeeds in fabricating 2-nanometer chips, how could it beat long-entrenched leading players on cost, customer ecosystem, and yield? Moreover, the customers for leading-edge logic chips are almost all clustered in AI accelerators and high-end processors—a domain where Japan lacks system-level strengths.

Two Paths, the Same Anxiety of the Era

From the perspective of industrial policy, Kumamoto and Rapidus embody two classic models: one secures certain dividends by integrating into the existing system through the global division of labor; the other pursues a technological leapfrog to build an independent and controllable commanding height. Japan's reality, however, is that it craves both, while its resources and risk tolerance are limited.

At the seminar, Akaba concluded that Japan's semiconductor strategy is essentially about balancing "domestic resilience" with "global integration." The Kumamoto project fortifies the resilience of the domestic supply chain, while Rapidus seeks to carve out a niche in the globalized supply chain. Yet economic security is neither equivalent to total self-sufficiency, nor to the blind pursuit of state-of-the-art process nodes. This ambition to have it both ways seems especially fragile in a fast-changing international environment.

Decisions Under the Shadow of Geopolitics

The US-China technology war has injected a deeper geopolitical calculus into Japan's semiconductor strategy. On the one hand, Japan wishes to reduce its reliance on a single source such as Taiwan; on the other, it must preserve deep industrial ties with the island. In a sense, the establishment of the Kumamoto plant is a typical exercise of "friend-shoring" within the US-led alliance system. Yet, should the Taiwan Strait situation turn volatile, the Kumamoto plant could fall victim to the fallout. Rapidus's quest for autonomy likewise signifies an attempt to hedge against geographical supply-chain risks.

But the danger is that Japan may be caught in a dilemma: over-dependence on foreign technology would undermine strategic autonomy, while an all-out bet on domestic breakthroughs could incur enormous sunk costs from technological barriers and market obstacles.

Future: An Experiment in "Resilience"In any case, these two projects have entered the implementation stage, and their success or failure will determine the bottom line of Japan's economic security. If the Kumamoto factory can continue to expand production and successfully drive the upgrade and iteration of domestic equipment and materials, Japan will gain a more stable base for mature process technology. And as long as Rapidus can cross the threshold of mass production, even if it cannot fully challenge TSMC, it will have the opportunity to establish a differentiated presence in specific advanced nodes.

From the perspective of the long cycle of technological evolution, semiconductors are not only an industry, but also a "heavy asset" of geopolitics. The core proposition of Japan's current game may not be how to return to the peak of the 1980s, but how to redefine its irreplaceability in the new global division of labor. However, even the most meticulous industrial policy cannot completely eliminate an uncertain future.

Reference source: A Tale of Two Approaches to Revitalize Japan's Semiconductor Industry