TSMC's 2nm process "N2" has entered a phase of expanding its mass production lines. Taiwan's Economic Daily News reported on July 27, 2026, citing industry sources, that Hsinchu's Fab 20 has reached a monthly output of 20,000 wafers, and that TSMC has commended the plant for this achievement. TSMC has not disclosed this monthly figure officially, but in Q2 2026, N2 accounted for 3% of wafer revenue, with rapid production increases expected in the following Q3. The 20,000-wafer-per-month figure serves as a gauge of Fab 20's ramp-up, but it does not represent TSMC's overall output or good-die count.
The 20,000-Wafer Figure Is a Reported Value for Fab 20 Alone
The 20,000 wafers per month reported by Economic Daily News represents Fab 20's production scale alone. The same newspaper also reported that N2 mass production ramp-ups are progressing across five plants in Taiwan. Meanwhile, TSMC's official pages identify Fab 20 and Kaohsiung's Fab 22 as N2 production sites, and its earnings briefings describe production increases across multiple phases in Hsinchu and Kaohsiung. Because the counting methods for plant buildings, sites, and manufacturing phases differ, the reported figure of five plants and the official designation of two sites cannot be directly compared.
Furthermore, it is unclear whether the reported "20,000 wafers per month" refers to monthly input wafers, processed wafers, or good-die-equivalent wafers. Nor does it represent TSMC's overall N2 capacity figure. Calculating the number of chips that can be manufactured or the resulting revenue from Fab 20's monthly output figure would require die area, yield, and product mix data.
Still, the report that a single plant has reached a scale of 20,000 wafers per month carries significance. TSMC had explained that it moved N2 into mass production with "good yields" in Q4 2025, and that by April 2026, production was increasing at both Hsinchu and Kaohsiung in response to demand from smartphones and high-performance computing (HPC)/AI. This latest figure indicates that the challenge has shifted from achieving the mass-production start date to expanding output at each plant.
3% Revenue Contribution Reflects the Current State Two Quarters After Mass Production Began
In Q2 2026, N2 contributed 3% to TSMC's wafer revenue for the first time. In the same quarter, N3 accounted for 30%, N5 for 33%, and N7 for 11%, with advanced technologies of N7 and below totaling 77%. N2 remains the smallest revenue-share advanced node, but it appeared in the revenue mix by its second quarter after mass production began.
The 3% revenue share and the reported figure of 20,000 wafers per month do not measure the same thing. There is a time lag between wafer input and its recognition as revenue in a customer's product, and pricing and product mix also differ by technology. TSMC's overall Q2 shipment volume was 4.336 million 12-inch equivalent wafers, but this is the total across all processes, not N2 shipment volume specifically.
The impact of the ramp-up also showed up in inventory. TSMC's days of inventory increased by 7 days from the previous quarter to 87 days, and the company attributed this primarily to N2's production increase. TSMC has disclosed only the relationship between production increases and rising inventory, without revealing the breakdown of N2 inventory or shipment timing. Rather than the sheer size of the monthly output figure, how much N2's revenue share grows in Q3 will serve as the gauge for measuring the effectiveness of the production ramp-up.
Managing the Switch to GAA Across Two Sites
N2 marks TSMC's first adoption of gate-all-around (GAA) nanosheet transistors in a mass-production process. While shifting the transistor structure away from the conventional FinFET, production is being expanded across two sites, Fab 20 and Fab 22. This is not a transition that can be accomplished simply by adding more manufacturing equipment; it requires a process that reproduces the new structure with stable yields.
According to TSMC's research materials, N2 offers up to 15% higher speed at the same power consumption compared to N3, or up to 30% lower power consumption at the same speed. Chip density exceeds 1.15x. This is why both AI servers, which want to increase processing performance within the same power envelope, and smartphones, which want to boost performance while preserving battery life, are candidates for adoption.
However, these performance figures are not measured values for individual products but comparative values for the process presented by TSMC. The benefits customers actually obtain will vary depending on circuit design and operating conditions. Even as monthly production capacity increases, N2's revenue will not grow at the same pace unless customer product qualification and design migration proceed accordingly.
Beyond the 3-4 Point Gross Margin Burden, Toward N2P
TSMC expects that N2's rapid production increase in Q3 2026 will lower overall company gross margin by approximately 3-4 percentage points. The gross margin forecast for Q3 is 65-67%, with a midpoint of 66%. Throughout the second half of 2026, the company has also factored in a 3-4 point decline stemming from N2's ramp-up. Whether N2's output growth translates into revenue growth, and when the burden on overall gross margin begins to shrink, will be the key profitability milestone.
Production capacity itself may expand faster than originally planned. At its 2026 technology symposium, TSMC presented a chart showing N2-family capacity growing at an average annual rate of approximately 70% from 2026 to 2028. When asked about this assumption at the July earnings call, CEO C.C. Wei responded that the current plan is "even bigger," though he did not disclose updated figures.
Following N2, N2P—which offers improved performance and power efficiency—is scheduled to enter mass production in the second half of 2026. Fab 20's 20,000 wafers per month is a waypoint. Whether N2's revenue share grows from Q3 onward, and whether the 3-4 point burden on gross margin begins to shrink, will be the benchmark for determining whether the early stage of the mass-production curve has been cleared.