Japan’s push into advanced semiconductor manufacturing just got a major boost. Rapidus, the country’s most aggressive foundry player, has revealed plans to ramp up 2nm production to 60,000 wafers per month by 2028—a capacity leap that would position it as a serious challenger to TSMC’s dominance. The move comes as the global chip industry grapples with soaring AI-driven demand, and Rapidus is betting that its 2HP process, with a claimed logic density of 237.31 MTr/mm², can compete directly with TSMC’s N2 node while offering a unique production edge.
The timeline is aggressive. By 2027, Rapidus expects to hit 6,000 wafers per month, then quadruple that output the following year. The company is also doubling down on single-wafer front-end processing, a method designed to fine-tune adjustments at low volumes before scaling improvements—a strategy that could help it avoid some of the pitfalls of traditional multi-wafer fabrication. But whether this approach translates into high-volume manufacturing (HVM) success remains unproven.
The 2nm Gambit
Rapidus isn’t just chasing TSMC’s footsteps; it’s aiming to leapfrog the competition. The 2HP process, if it meets expectations, would offer performance and efficiency comparable to TSMC’s N2, while potentially undercutting Intel’s 18A node in density. Yet, the foundry’s roadmap extends even further: construction on a 1.4nm facility is slated to begin in 2027, with whispers of 1nm wafers in the pipeline—though no concrete details have been confirmed.
What sets Rapidus apart is its focus on Japan’s domestic market and a strategy to attract major U.S. customers, including those in defense and AI. The company has already secured early adopters for its 2nm process, signaling demand exists—but whether it can sustain that momentum at scale is another question. TSMC, meanwhile, is expanding its own 3nm production in Kumamoto, a move that could indirectly pressure Rapidus to accelerate its timeline.
Unanswered Questions
The biggest unknown is whether Rapidus’s 2HP node can deliver on its density promises in real-world production. TSMC’s N2 has already faced delays, and Rapidus’s single-wafer approach, while innovative, introduces risks in yield and cost. Additionally, the foundry’s 2028 target for full-scale 2nm production is optimistic given the complexity of ramping advanced nodes.
There’s also the elephant in the room: TSMC’s Kumamoto fab, now confirmed to produce 3nm chips, could absorb some of the demand Rapidus hopes to capture. If TSMC’s expansion succeeds, Rapidus may find itself playing catch-up in a market where first-mover advantage matters most.
A Foundry Showdown in the Making
The semiconductor landscape is shifting. Rapidus’s push into 2nm isn’t just about technology—it’s about geopolitics. Japan’s government has heavily invested in the foundry to reduce reliance on TSMC and U.S. suppliers, and Rapidus’s ambitions align with broader national goals. But without proven HVM success, even the most aggressive roadmap could falter.
One thing is clear: the race for 2nm and beyond is heating up. Whether Rapidus can turn its plans into reality will determine whether Japan’s semiconductor renaissance becomes a reality—or just another ambitious promise left unfulfilled.
