Japan spent another 944 million US dollars to strengthen the “chip national team” Rapidus, and the 2-nanometer process circuit accelerates to catch up with TSMC

The Zhitong Finance App learned that in the context of the intensification of the global semiconductor competition, the Japanese government is once again adding to its “semiconductor national team.” Japan's Ministry of Economy, Trade and Industry announced that it plans to allocate 150 billion yen (approximately US$944 million) of additional investment capital to advanced semiconductor manufacturer Rapidus Corp. in the 2027 fiscal year budget to increase spending and competitiveness with leading companies such as TSMC.

Following the Japanese government's investment of 100 billion yen in February and approval of 631.5 billion yen in R&D support in April, this is another key injection of capital by the Japanese government into this startup known as the “Japanese version of TSMC.” If this additional investment of 150 billion yen is calculated by combining existing subsidies, the Japanese government and private capital support for this state-owned chip company founded in 2022 is approaching the scale of 3 trillion yen by 2027.

The “chip breakout war” under the national system

Rapidus was founded in August 2022 and was jointly funded by eight Japanese giants including Toyota, Sony, SoftBank, Kioxia, Nippon Denso, NEC, Nippon Telecom Telephone, and Mitsubishi UFJ Bank. As the result of collaboration between Japanese industry, government, and academia, this company carries the core mission of reviving the Japanese semiconductor industry — to achieve domestic mass production of 2nm cutting-edge chips by 2027, breaking excessive dependence on overseas foundries such as TSMC.

The Japanese government has elevated the success of Rapidus to the level of a national security strategy. Policymakers believe that mastering autonomous chip manufacturing capabilities in the fields of artificial intelligence, robotics, and quantum computing is essential to Japan's technological sovereignty. To ensure that core technology is not exported, the Japanese government also holds “gold shares” of Rapidus and has veto power over important matters.

Chaser Rapidus timeline and price war

Rapidus' mass production schedule has been clearly defined: 2 nm chip production will begin in the second half of FY2027, and full mass production will be achieved in FY2028. Within about a year after mass production starts, the monthly production capacity target will be raised to about 25,000 wafers, helping Japan get rid of its dependence on TSMC. Japanese policymakers regard Rapidus' success and technological independence in the fields of AI, robotics, and quantum computing as a strategic key to national security.

At the technical level, Rapidus has obtained IBM's 2nm process technology license and is equipped with ASML's high-numerical aperture extreme ultraviolet (high-NA EUV) lithography equipment. In July 2025, the company successfully developed a 2 nm full surround gate (GAA) transistor that works properly. The trial production process only took 1/3 to 1/4 of the time of TSMC or Samsung's conventional process.

In terms of pricing strategy, Rapidus plans to provide foundry services at a price of 3 million to 3.5 million yen (approximately 18,500 to 21,500 thousand US dollars) per wafer. This price is nearly 10,000 US dollars cheaper than TSMC's price of about 30,000 US dollars for 2 nanometer wafers, and the price reduction is as high as 30%. Rapidus CEO Atsuyoshi Koike said bluntly that as a latecomer in the market, the company “cannot lose to TSMC” in terms of pricing.

However, the road to catch up is still difficult. TSMC began mass production of 2 nanometer chips in 2025, and its production capacity is rapidly expanding. The monthly production capacity is expected to reach 70,000 to 75,000 chips by the end of 2026. In contrast, Rapidus started one step later, and the mass production scale was only about one-third that of TSMC.

Triple competitive clash: TSMC, Samsung, and Musk

The competitive landscape that Rapidus faces is far more complex than expected.

TSMC is the biggest rival. Its 2 nm production capacity has been fully booked by customers such as Apple and Qualcomm this year. The four factories have a monthly production capacity of 60,000 tablets, and the trial production yield is over 70%. TSMC's first-mover advantage in technology accumulation and customer ecology is difficult to shake in the short term.

TSMC began mass production of 2 nm in 2025. Its CoWoS advanced packaging roadmap is accelerating from 5.5 times the mask size in 2026 to 14 times that of 2028. In 2026, TSMC plans to invest more than $50 billion in capital expenditure — a figure that far exceeds all the resources that Rapidus can mobilize.

Samsung Electronics is also eyeing the 2nm field and is in direct competition with Rapidus.

What's even more dramatic is that Elon Musk is joining forces with the “Terafab” project. Musk is cooperating with Intel to advance this giant chip project, with a target annual production capacity of 1 terawatt of computing power to manufacture semiconductors for companies such as Tesla, SpaceX, and XAI. Terafab plans to invest up to 5 trillion US dollars, and its technology route is in direct conflict with Rapidus.

Notably, Rapidus' 2nm technology is licensed by IBM, and Musk's Terafab chose Intel as a partner. This means that on a 2nm and more advanced circuit, Rapidus is playing against a “super league” composed of Musk+Intel.

The trillion gap and the test of commercialization

Despite continued blood transfusions by the Japanese government, Rapidus' funding gap is still huge. By 2031, Rapidus plans to invest more than 3 trillion yen in the development and mass production of 1.4 nm and 1 nm semiconductors, and the total investment will expand to over 7 trillion yen. At present, the Japanese government has decided to provide a total of about 2.9 trillion yen in support funds, and the company still needs to raise about 4 trillion yen.

Rapidus has yet to achieve commercial mass production. It will complete a 2 nm GAA test chip flow sheet in 2025, and deliver a 2 nm process design kit (PDK) to customers in the first quarter of 2026. From trial production to mass production, from technical verification to customer acquisition, Rapidus' commercialization path is still full of variables.

What is unavoidable is the real pressure on energy and material costs. In resource-poor Japan, Rapidus is being hit by rising energy and material input costs due to the Middle East conflict.

The schedule for mass production is extremely tight. Judging from the goal of mass production of 2 nm in 2027, Rapidus needs to complete the leap from trial production to large-scale mass production in less than two years, which is unprecedented in the history of the semiconductor industry. Korean scholars have questioned that its “single-wafer processing model is simply not cost effective.”

What's more critical is customer acquisition. TSMC has firmly locked in core customers such as Nvidia and Apple. To this end, the Japanese government is subsidizing Fujitsu and IBM in Japan to develop AI semiconductors and push them to commission Rapidus foundry in an attempt to establish an “anchor customer” for them before mass production.