Rapidus Is the Last Chance for Japan to Matter in Semiconductors

Time is running out as veteran engineers retire, says expert Kazuto Suzuki, and Rapidus must turn their know-how into working chips in time. 

Japan has committed roughly $15 billion in government support to Rapidus, a semiconductor manufacturer established in 2022, in an ambitious effort to revive the country's advanced chipmaking industry. The company aims to begin mass production of cutting-edge 2-nanometer chips in the second half of fiscal 2027.

Rapidus has already made progress toward that goal, including opening a pilot production line at its Hokkaido factory and demonstrating its ability to produce prototype 2-nanometer transistors. 

However, manufacturing chips on a commercial scale is a much greater challenge. The company must prove it can consistently produce large quantities of working chips at a competitive cost and convince customers to order.

Kazuto Suzuki, a professor of science and technology policy at the University of Tokyo's Graduate School of Public Policy and director of the Institute of Geoeconomics, believes Japan's investment is justified. 

However, speaking to Japan Forward, he emphasized that Rapidus should not be viewed primarily as a safeguard against a potential conflict over Taiwan.

Two Halves of One Strategy

Taiwan Semiconductor Manufacturing Company (TSMC), the world's largest contract chipmaker, has also been expanding its presence in Japan. Its first factory in Kumamoto Prefecture began mass production in late 2024, while a second is under development.

Suzuki sees TSMC's investment and Rapidus as serving two different purposes in Japan's economic security strategy.

Rapidus aims to make Japan an indispensable part of the global semiconductor supply chain by producing advanced chips. TSMC's factories, meanwhile, are intended to strengthen Japan's ability to meet domestic demand without relying entirely on imports.

Japan once dominated the semiconductor industry, accounting for roughly 50% of the global market in the 1980s. Today, its share is below 10%, and Suzuki believes time is running out to reverse the decline.

"I think this is going to be the last opportunity for Japan to take a gamble to be a meaningful player in the global market," he said.

The biggest concern is the loss of expertise. Engineers who helped build Japan's semiconductor industry are approaching retirement, taking decades of practical knowledge with them. Suzuki argues that younger engineers trained overseas cannot easily replace that experience.

"In the semiconductor world, experience is everything," he said.

That experience is particularly important for improving production yields, the percentage of chips manufactured on a silicon wafer that function properly. 

Because wafers are expensive to produce, even small yield improvements can make a major difference to profitability. Suzuki praised the government's decision to bring in Atsuyoshi Koike, now Rapidus CEO, to lead the effort to rebuild Japan's chipmaking capabilities.

Two-Pronged Semiconductor Strategy 

Suzuki said Rapidus could eventually rival TSMC's technological capabilities but operates under a different business model and isn't intended to replace it.

Japan's major chip buyers, including Sony, Denso, and Toyota, have limited demand for the cutting-edge processors Rapidus plans to manufacture. 

Unlike companies producing AI servers and smartphone processors, they generally rely on less advanced chips. Rapidus will therefore depend largely on overseas customers.

The first plant began mass production in late 2024, producing 12- to 28-nanometer chips used in automobiles and industrial equipment.

TSMC is also expanding its advanced chipmaking capabilities in Japan. Its second Kumamoto plant is now planned to produce 3-nanometer chips, with operations expected to begin in 2028. 

The Japanese government has approved subsidies of up to ¥732 billion ($4.7 billion) for the project, which Prime Minister Sanae Takaichi has described as significant for economic security.

Who Will Fill the Fab?

Suzuki is optimistic about Rapidus finding customers. With demand for advanced chips surging because of AI and supplies struggling to keep pace, he believes "One way or another, there'll be customers from all over the world."

He also sees mass production as a necessary step, not a major obstacle. Private investors, he explained, are largely waiting for Rapidus to demonstrate its manufacturing capabilities before committing more funds.

In February, Rapidus raised ¥267.6 billion ($1.7 billion), including ¥167.6 billion from 32 private companies, exceeding its ¥130 billion target. CEO Koike has said the company is in talks with more than 60 potential customers.

However, questions remain about whether Rapidus can secure enough customers to make its business profitable. On October 5, the company announced partnerships with 17 firms, including Synopsys and Infosys, to help customers design chips. While this marks progress, it does not necessarily translate into orders.

Convincing customers to switch from established manufacturers such as TSMC will also be difficult, particularly given Rapidus's lack of a commercial production track record. Still, the company could attract smaller customers whose orders larger manufacturers prioritize less.

Success Is Measured in Yield

Suzuki offers a simple test for the whole policy. "If you make a good chip, then that's a success," he said. "If you improve the yield, that's a success."

Rapidus reached an important milestone in July 2025 when its prototype 2-nanometer transistors demonstrated the expected electrical performance. However, the company has not yet disclosed its production yield, a key measure of how many chips it can manufacture successfully.

Consistently producing advanced chips at a commercially viable cost remains a major challenge. Rapidus aims to begin mass production in fiscal 2027 and is targeting a stock market listing around fiscal 2031.

Suzuki believes it is too early to judge the company's prospects, with mass production providing the first meaningful test.

"It's still wait and see," he said. "But so far, Rapidus has been on schedule, and this is important."

RELATED:

Rapidus: Japan's Return to the Frontier of Semiconductor Manufacturing

Author: Daniel Manning