Rapidus Puts a Price on 2nm: ¥3–3.5 Million per Wafer — Undercutting TSMC Before Shipping a Single Production Chip

In the foundry business — contract chip manufacturing — price lists normally stay hidden. At the leading-edge 2nm generation, TSMC is effectively the only company in volume production, and the power to set prices has rested with TSMC. Now a Japanese newcomer with zero production track record has shown a price tag cheaper than the incumbent's, before shipping a single production chip. That is what makes this report remarkable.

What was said

Here is a summary of the Nikkei report (July 8, 2026) and the TrendForce roundup of it (July 9, 2026). The remarks were made at a Japan Productivity Center event held in Karuizawa on July 8, 2026.

One caveat matters here. These are remarks from a CEO speech as reported by the press, not an official Rapidus press release. Every figure in this article should therefore be read as media-reported.

Why does a latecomer with zero track record lead with price?

What foundry customers fear most is entrusting a design to a fab that cannot manufacture it properly. TSMC has decades of volume-production history; Rapidus has yet to ship a single production wafer. For a latecomer to give customers a reason to switch, it needs something other than track record — and the most legible answer is price.

The timing is also favorable. TSMC has reportedly notified customers of price increases across all of its leading-edge nodes, according to Culpium, the independent outlet run by former Bloomberg journalist Tim Culpan (June 23, 2026; the size of the increases was not disclosed). When the incumbent raises prices, the challenger declares it will match or undercut. The wider that gap grows, the more easily those ongoing talks with 60-plus companies turn into concrete order evaluations.

How Rapidus intends to hit this price, on the other hand, has not been disclosed in the reporting. What is clear is the financial structure underneath it. Rapidus has received a cumulative 2.354 trillion yen in government funding (as of April 2026), with the state shouldering a large share of the massive upfront investment. The history of that support and the remaining walls before mass production are covered in Japan Has Committed ¥2.35 Trillion to Rapidus — and Money Was Never the Hard Part. The company's differentiation strategy of single-wafer processing and short TAT (shortened turnaround time) is covered in Rapidus Explained: Japan's Last Bet on a 2nm Foundry.

What does the 2nm market look like right now?

Behind the price declaration is a severe 2nm supply shortage. TSMC's 2nm (N2) only entered volume production in December 2025, and its 2026 capacity is reported as effectively sold out to Apple, NVIDIA, Qualcomm and AMD. Anyone who wants the leading edge is waiting in line, so structural demand for a second source exists. The positions of all four contenders — TSMC, Intel, Samsung and Rapidus — are mapped in 2nm was sold out before it was built: where TSMC, Intel, Samsung and Rapidus stand (in Japanese).

Rapidus's own roadmap is moving as well. A 2nm pilot line is already running at IIM-1 in Chitose, Hokkaido, with volume production targeted for the second half of 2027. Beyond that, the 1.4nm generation is slated for development start in 2026, fab groundbreaking in 2027, and volume production in 2029. In other words, this price declaration is not a one-off discount — it is the number placed at the entrance of a multi-generation strategy: win customers at 2nm, keep them at 1.4nm.

How should investors read this?

Rapidus is unlisted, so direct investment is not possible. But the pricing strategy carries real signals for the market. First, a price in the 3-million-yen range per wafer is a statement that share capture takes priority over profit. A foundry almost never starts its ramp in the black, and near-term economics should be read as designed around continued state support.

Second, the milestones to watch are now clear: whether the 60-plus company talks convert into the announcement of a first volume-production customer, and whether the second-half-2027 production start happens on schedule. These two points are the dividing line between a price declaration that materializes and one that remains on paper. For Japan's equipment and materials suppliers, Rapidus reaching volume production is likewise the checkpoint that determines whether incremental demand appears.

What are the caveats?

The biggest caveat is that this is not a current selling price. The 3–3.5 million yen figure is a target for the production start in the second half of 2027 — a goal more than a year out. If exchange rates or equipment and materials costs move, the premises move with them.

Next, yield — the share of good chips obtained from each wafer — is unproven. A customer's effective cost is wafer price divided by yield, so even a 30% cheaper wafer is not cheap for the customer if yield badly trails the incumbent's. At the same time, taking orders at low prices with poor yield means Rapidus digs its own losses. Cheapness only becomes a weapon once volume-production capability backs it up.

Finally, consider the nature of the source. The price came from a CEO's remarks at a speech, not an officially committed rate card. Even so, in an industry where the norm is to say nothing about price at all, the fact that a latecomer deliberately showed concrete numbers first carries weight. The challenger is compensating for its lack of track record with price transparency — that is the most accurate way to read this report (Nikkei, TrendForce).