--- title: "Japan earmarks another $944 million for Rapidus, its long-shot bet against TSMC" description: "Tokyo is setting aside a further ¥150 billion, about $944 million, for Rapidus, the state-backed chipmaker trying to build 2-nanometre logic capacity in Hokkaido from a standing start. The Ministry of Economy, Trade and Industry will seek more in its fiscal 2027 budget, on top of public support already running into the trillions of yen." category: "Tech" category_url: https://boursel.com/category/tech author: "Daniel Okonkwo" published: 2026-08-21T13:47:00.000Z updated: 2026-08-21T13:47:00.000Z canonical: https://boursel.com/article/japan-earmarks-another-944-million-dollars-for-rapidus-its-long-shot-bet-against tags: ["semiconductors", "japan", "rapidus", "industrial-policy", "tsmc"] --- # Japan earmarks another $944 million for Rapidus, its long-shot bet against TSMC Tokyo is setting aside a further ¥150 billion, about $944 million, for Rapidus, the state-backed chipmaker trying to build 2-nanometre logic capacity in Hokkaido from a standing start. The Ministry of Economy, Trade and Industry will seek more in its fiscal 2027 budget, on top of public support already running into the trillions of yen. Japan is putting another ¥150 billion, about $944 million, behind Rapidus, [Bloomberg reported](https://www.bloomberg.com/news/articles/2026-08-21/japan-earmarks-another-944-million-for-rapidus-in-ai-chip-race), citing the Nikkei. The Ministry of Economy, Trade and Industry will ask for further funds in its fiscal 2027 budget request. By itself the figure is unremarkable next to what has already gone in. What it signals is that the government intends to keep funding the project through the most expensive stretch of its plan, the ramp from a pilot line to volume production. ## What Rapidus is Rapidus was created in 2022 as a joint venture between the Japanese government and a group of Japanese companies, to do something Japan has not done in thirty years: manufacture leading-edge logic chips at home. Its fab is at Chitose in Hokkaido, its process technology comes from a partnership with IBM, and it has drawn on Belgium's imec for prototyping and extreme ultraviolet lithography expertise. Its private shareholders read like a roll call of corporate Japan. In a ¥267.6 billion round in February, [the state's Information-technology Promotion Agency put in ¥100 billion and 32 companies the remaining ¥167.6 billion](https://www.rapidus.inc/en/news_topics/information/rapidus-secures-267-6-billion-yen-in-funding-from-japan-government-and-private-sector-companies/), among them Canon, Sony, SoftBank, Toyota, Honda, NTT, Fujitsu, NEC, Kioxia and the Development Bank of Japan. Public support is on another scale entirely. Cumulative government backing, counting subsidies, research commissions and equity, runs into the trillions of yen and is projected to reach the equivalent of tens of billions of dollars by the end of the current fiscal planning horizon. This is industrial policy of the kind Japan largely abandoned after the 1990s. ## The bet The target is mass production of 2-nanometre logic chips from the second half of fiscal 2027, starting at a few thousand wafers a month and scaling from there. Rapidus has run test wafers through a gate-all-around process and reported working transistors. The difficulty is hard to overstate. Japan [held more than half of world semiconductor production in the late 1980s and holds under 10% today](https://www.csis.org/analysis/japan-seeks-revitalize-its-semiconductor-industry), and it has no recent experience at the leading edge. Rapidus is attempting to skip several process generations at once, which is not how anyone else got there: TSMC, Samsung and Intel each climbed node by node, learning yield at every step. Yield is the whole game. A leading-edge fab is economic only when a high enough share of the chips on each wafer works, and that share is won through years of iteration on a running line. A company with no production history is starting that clock late, against competitors already shipping 2nm. ## What Japan is actually buying The commercial case for Rapidus as a challenger to TSMC is weak, and most analysts say so. The strategic case is different and better. Japan remains indispensable further up the supply chain, with dominant positions in photomasks, photoresists and semiconductor manufacturing equipment. What it lacks is the ability to fabricate an advanced chip inside its own borders if Taiwan became unavailable. That is the risk the spending addresses, and it is not one a discounted cash flow model prices. The plausible commercial niche is narrower than the headlines suggest: short runs with fast turnaround for customers who want a supplier outside Taiwan and are willing to pay for it, rather than volume competition on price. ## What to watch Three things will tell you whether this is working before the 2027 date arrives. Whether Rapidus names committed customers rather than discussions. Whether it publishes yield data or continues to describe milestones qualitatively. And whether the private shareholders put in more money at the next round, or leave the government to carry it alone. The last of those is the most informative. Corporate Japan has been willing to fund this so far as a national project. Whether it funds it as an investment is a different question. ## Sources - [Japan earmarks another $944 million for Rapidus in AI chip race](https://www.bloomberg.com/news/articles/2026-08-21/japan-earmarks-another-944-million-for-rapidus-in-ai-chip-race) - [Rapidus secures 267.6 billion yen in funding from Japan government and private sector companies](https://www.rapidus.inc/en/news_topics/information/rapidus-secures-267-6-billion-yen-in-funding-from-japan-government-and-private-sector-companies/) - [Japan's Rapidus receives another ¥267.6 billion in funding](https://www.japantimes.co.jp/business/2026/02/27/companies/rapidus-new-investment/) - [Japan seeks to revitalize its semiconductor industry](https://www.csis.org/analysis/japan-seeks-revitalize-its-semiconductor-industry)