There is a question worth asking before celebrating Japan's latest semiconductor ambition: when a government becomes your biggest shareholder and reserves the right to veto your management decisions, is it still a private enterprise — or is it industrial policy wearing a startup's clothing?
The answer, in the case of Rapidus Corporation, is probably both. And that ambiguity sits at the heart of what is either one of the most strategically shrewd moves in modern industrial policy or one of the most expensive gambles in semiconductor history.
On 27 February, Rapidus announced it had closed a funding round totalling ¥267.6 billion, equivalent to approximately $1.7 billion Australian dollars, drawn from the Japanese government and 32 private-sector companies. ¥100 billion of the funding came from the Information-technology Promotion Agency (IPA), an independent administrative body under the jurisdiction of the Ministry of Economy, Trade and Industry. The remaining ¥167.6 billion was sourced from 32 companies including Canon, Fujitsu, NTT, SoftBank, the Development Bank of Japan, and Sony Group.
According to Japanese newspaper The Mainichi, the Japanese government has now become the largest shareholder in Rapidus, holding 11.5 per cent of its voting rights, as well as a "golden share" giving it veto power over key management decisions. That is a striking degree of state intervention for a venture that still frames itself as a commercial foundry. Defenders of industrial policy will say this is precisely the kind of strategic commitment needed to challenge an entrenched duopoly. Sceptics will note that governments rarely make patient, disciplined investors.
A Mountain to Climb
The scale of the challenge facing Rapidus cannot be overstated. This is an ambitious plan for a four-year-old firm; Rapidus could position itself as a rival to chipmaking giants TSMC and Samsung, both of which started volume production using their own 2nm silicon processes in Q4 of 2025. That head start matters enormously. All of TSMC's 2nm production capacity at its two dedicated fabs for 2026 has already been booked, with Apple accounting for more than half of the initial capacity. Getting a foot in the door when the market leaders have already locked up their order books requires a genuinely differentiated pitch.
This latest round of government funding comes on the back of real technical progress. Rapidus opened its pilot line in April 2025 and demonstrated working 2nm gate-all-around (GAA) transistors — an architecture that surrounds the transistor channel on all four sides rather than three, improving current control and reducing leakage compared to FinFET designs — shortly after, in July. That is not a trivial achievement for a company that did not exist four years ago.
The fab is being built in Chitose, Hokkaido. Rapidus was founded in 2022 with the aim of revitalising Japan's fading semiconductor industry, and has signed up help from IBM and Belgian-based R&D organisation Imec (Interuniversity Microelectronics Centre) for the development of the technology required to reach its goal. That international partnership structure is the project's most credible technical foundation. IBM's deep knowledge of GAA transistor architecture, combined with Imec's process research capability, gives Rapidus access to intellectual capital it could not have built domestically in this timeframe.
The Counter-Argument Deserves Serious Consideration
Critics of state-directed semiconductor investment point to a well-worn history of failure. National technology champions have been built and abandoned across Europe and Asia for decades. The foundry business demands brutal capital discipline and years of yield improvement before any customer will trust you with their most sensitive designs. Governments, the argument goes, are constitutionally incapable of that kind of patient, dispassionate management.
But strip away the ideological reflexes and what remains is a more complex picture. The semiconductor supply chain is not a normal market. As Omdia principal analyst Manoj Sukumaran told The Register, the strategic case is compelling from a sovereignty standpoint: "It's a Japanese initiative and the intention is to regain Japan's lost leadership in leading-edge semiconductor manufacturing." That framing resonates well beyond Japan. Australia, a country deeply enmeshed in US-aligned supply chains that run through Taiwan, has a direct interest in whether the Indo-Pacific region can develop genuine semiconductor redundancy.
Sukumaran also made the commercial case clearly: "The value proposition of Rapidus is small volumes at fast turnaround time. Being a small fab and also aligning the manufacturing setup for a large number of small batches will help to attract companies who are iterating their chip designs fast." For AI startups, defence contractors, and specialist chip designers who cannot secure capacity at TSMC and do not want Samsung's current yield challenges, that proposition is genuinely attractive. Unlike the industry-standard batch processing model, which can take up to 120 days to cycle a wafer through a fab, Rapidus is utilising a proprietary single-wafer processing system aimed at slashing cycle times to just 50 days.
What the Money Signals
Rapidus is in talks with more than 60 companies eager to design chips for artificial intelligence, robotics, and edge computing, according to CEO Atsuyoshi Koike, who spoke at a news conference in late February. That pipeline, if real, suggests the commercial logic is not entirely dependent on government goodwill.
The involvement of Japan's largest industrial names carries its own signal. Sony and SoftBank are not sentimental investors. Canon and Fujitsu have direct competitive interests in advanced semiconductor access. Fujifilm, a supplier of semiconductor raw materials contributing to the latest investment round, stated it will support Rapidus by providing a range of technologies and will work closely with it on next-generation process development. This is a consortium of companies betting that the cost of dependence on foreign foundries now exceeds the risk of backing a domestic alternative.
The fab will also have to integrate with electronic design tools, and chipmakers will have to test and validate their intellectual property on the new process node before it can deliver any products, so there is a long way to go yet. The timeline from pilot line to mass production is where most semiconductor ambitions have historically collapsed. Yield rates, customer qualification cycles, and equipment integration failures have derailed better-funded efforts than this one.
The fundamental question is not whether Japan should have a domestic semiconductor industry. The answer to that, on national interest grounds, is almost certainly yes. The harder question is whether a state-backed foundry — with a government golden share and a 2027 deadline — is the right instrument to get there. History suggests caution. The technical milestones to date suggest the sceptics should not be too comfortable either. Reasonable people can disagree on whether this is strategic genius or expensive optimism; what they cannot disagree on is that the outcome will matter, well beyond Japan's borders.