How Japan-based Rapidus is trying to become a globally competitive chipmaker by 2027, aiming to mass produce 2nm chips just two years after TSMC and Samsung
Tetsuro Higashi is taking on what seems like an impossible task: Create a globally competitive semiconductor manufacturer in Japan from scratch - and do it in four years. Tweets: @gavinsbaker , @rodolfor , @kanthan2030 , @6d6f636869 , @pelstrom , and @pelstrom Tweets: Gavin Baker / @gavinsbaker : Important to watch - might actually work. All the equipment manufacturers doing everything they can to help TSM competitors and Moore's law slowing down makes it easier to catch up. Only possible to catch TSM in Japan, America, South Korea, Israel IMO. https://www.bloomberg.com/... Rodolfo Rosini / @rodolfor : This is the most Japanese thing possible: - catching up on technology others already have - CEO is 73yo - refuses to raise externally instead govt + incumbents round https://www.bloomberg.com/... S.L. Kanthan / @kanthan2030 : A new Japanese company wants to be a semiconductor leader within five years, rivalling TSMC and Samsung. Rapidus is mostly funded by the Japanese government, led by the former CEO of Tokyo Electron, and has partnerships with US/EU firms. Why not? https://www.bloomberg.com/... Takashi Mochizuki / @6d6f636869 : Out: Japan's effort to build its own TSMC from scratch — in four years. https://www.bloomberg.com/... Peter Elstrom / @pelstrom : Rapidus' success would mean the domestic ecosystem could sustain profits at a high enough level to remain relevant amid intensifying global competition. “Huge opportunities are ahead of us if we can become first in the market,” Higashi said. https://www.bloomberg.com/... Peter Elstrom / @pelstrom : Meet the man who is taking on what seems like an impossible task — building a globally competitive semiconductor manufacturer in Japan from scratch. Oh, and he's supposed to do it in four years. The story from @6d6f636869 w/Yuki Furukawa https://www.bloomberg.com/...
Context & Ripple Effects
When Bloomberg profiled Tetsuro Higashi's attempt to stand up a globally competitive chipmaker in Japan within four years, the bet looked quixotic: a company built from scratch targeting 2nm mass production by 2027, just two years after TSMC and Samsung. Higashi — formerly of Tokyo Electron — deliberately refuses outside capital, funding Rapidus instead through the Japanese government and incumbent industry firms.
The arc since has been one of escalating state commitment layered onto that structure: Japan's $67B domestic chip revival program made Rapidus its flagship, Atsuyoshi Koike took over as CEO with a ~$35B spend planned by 2027 to help the US counter China, and Tokyo has since topped up the war chest with a ~$5.4B aid tranche followed by a further ~$1.6B across two fiscal years. The strategic question has shifted from whether Japan will pay to what business model can make a late entrant viable.
First-order effects
- Japanese taxpayers and incumbent industry firms absorb nearly all of Rapidus's capital risk — successive government tranches plus a ~$696M share issuance give the venture runway without dilution to external investors Higashi refuses to court.
- TSMC and Samsung now face a state-backed entrant aiming at their most advanced node, even if it arrives roughly two years behind them.
Second-order effects
- Equipment suppliers have a structural reason to help a TSMC alternative exist — Gavin Baker's observation in the piece that toolmakers are 'doing everything they can' to assist TSM competitors aligns with Higashi's own Tokyo Electron roots, easing Rapidus's tooling access while eroding TSMC's ecosystem moat.
- The FT's reporting on Rapidus pushing efficient small-batch production of bespoke chips points to a second front: competing not on TSMC's volume economics but on low-quantity, high-mix orders that leading-edge incumbents price poorly.
Third-order effects
- If Moore's-law deceleration really does compress how far ahead the leaders stay — Baker's argument that catching up is only feasible in a handful of allied countries — then state-funded challengers like Rapidus could permanently fragment leading-edge foundry capacity away from the Taiwan-Korea duopoly, with governments rather than markets selecting who gets a seat.
- A viable Rapidus would make allied-government co-funding a standing feature of advanced-node competition, turning every future node transition into a geopolitical subsidy contest between Washington, Tokyo, Brussels and Beijing.
The trend: Leading-edge chipmaking is shifting from a market-selected duopoly toward state-capitalized regional champions, with Moore's-law slowdown narrowing the technical gap faster than the funding gap.