In a concerted effort to reclaim its historical dominance in semiconductor manufacturing, Japan is launching an ambitious project named Rapidus, designed to create a leading-edge semiconductor manufacturer. The initiative, which aims to overcome a significant competitive landscape primarily dominated by Taiwan, South Korea, and the United States, has been in development since its conceptualisation was presented to the then Prime Minister Shinzo Abe in 2020.
Atsuyoshi Koike, a veteran in the semiconductor industry and now Rapidus's chief executive, insists that this venture marks one of the most impactful technological pursuits for Japan since the Meiji Restoration. Currently, Rapidus is a highly capital-intensive start-up that aims to revolutionise the traditional model of semiconductor production which typically operates on a mass scale. Koike's vision proposes that custom chips can be beneficially produced in smaller, bespoke quantities, a divergence from the industry standard primarily exemplified by the Taiwan Semiconductor Manufacturing Company (TSMC).
"Ever since TSMC was founded by Morris Chang in 1987, it has been all about scale and Rapidus is really something nobody has ever attempted before," says David Dai, a semiconductor analyst at Bernstein, highlighting the potential game-changing impact of this venture should it succeed.
The factory is currently under construction in Hokkaido, roughly 900km from Rapidus's Tokyo headquarters. The facility is set to receive an advanced extreme ultraviolet lithography machine from the Dutch equipment manufacturer ASML, a crucial component in the production of target two-nanometre chips slated for trial production in April 2024. Mass production is targeted to commence by 2027, contingent upon the successful implementation of the technology.
Despite the ambitious timelines, skepticism looms regarding the viability of Rapidus's approach, particularly its intent to advance directly to the production of 2nm chips. Industry experts point out that this leap may be untenable, noting that the sheer costs associated with high-end chips produced in small quantities could deter potential customers. Konrad Young, a retired professor from TSMC, remarked, "I don’t think Nvidia will even bother. The cost will be much higher."
The Japanese government has already committed a significant financial investment, earmarking ¥920bn for the project and unveiling a broader package of ¥10tn ($65bn) aimed at the artificial intelligence and semiconductor sectors over the next seven years. Yoshihiro Seki, a senior figure in the ruling Liberal Democratic Party, remarked on the unprecedented level of focus the government has placed on semiconductor technology, acknowledging the colossal risks entailed in placing such a substantial effort on a single venture.
Historically, Japan enjoyed a commanding position in the semiconductor industry, holding more than 50 per cent market share at the end of the 1980s. This dominance has waned, with Japan's current share declining to approximately 15 per cent, largely due to an underestimation of competitor movements and an inability to adapt collaboratively within the industry.
Nevertheless, geopolitical factors today compel Japan to bolster its semiconductor capabilities, particularly as tensions surrounding supply chains intensify and depend heavily on TSMC. The Japanese Ministry of Economy, Trade and Industry (Meti) is fostering an environment conducive to developing a semiconductor ecosystem that includes international partnerships, with Rapidus positioned as the cornerstone of this undertaking.
Despite facing numerous challenges, including a potential shortage of skilled labour and the distance from Japan’s educational resources, Koike continues to assert confidence in Rapidus's ability to innovate and succeed, stating that their unique approach, including a commitment to single-wafer processing, will yield significant results in chip manufacturing efficiency and yield rates.
However, the daunting reality of immense funding requirements contrasts sharply with their projected timelines. Analysts have elaborated that the escalating costs of developing each new chip generation—from roughly $1bn at 28nm to approximately $6bn at 3nm—pose considerable risks. With only 20 per cent of the projected ¥5tn funding secured, doubts persist over Rapidus's capacity to meet the ambitious production timeline before advancing technologies render them obsolete.
As Japan invests heavily in this initiative, the scrutiny on Rapidus will mount; some industry experts are already questioning whether the country can indeed regain its former prominence in semiconductor manufacturing. Regardless of the outcome, the establishment of a fabrication plant in Hokkaido will position Japan strategically in the global semiconductor landscape, indicating that its long-term ambitions extend well beyond the fate of a single company.
Source: Noah Wire Services