[News] Earnings Call Preview: TSMC 2nm in Focus — Capacity, Sales Contribution, and More

Ahead of TSMC’s earnings call on January 15, the company quietly confirmed that its 2nm process entered mass production on schedule in 4Q25. With the market now expecting more details at the briefing, here are some of the key things to watch as this long-anticipated node steps into the spotlight.

Monthly Capacity Could Challenge 140K

Commercial Times, citing analysts, notes that TSMC is rapidly ramping 2nm capacity at its Kaohsiung Fab 22 and the Baoshan fab, with monthly output expected to climb toward 100,000 wafers by the end of this year to meet demand from major customers including Apple, AMD, MediaTek, and Qualcomm. Liberty Times, being more optimistic, suggests its monthly 2nm output could reach 140,000 wafers, marking an extraordinary ramp-up just one year into production.

In comparison, the Liberty Times notes that TSMC’s 3nm output is set to expand to around 160,000 wafers per month this year, underscoring the sustained strength of demand for advanced process nodes.

Revenue Contribution to Start in Q3

However, the company may have to wait for a little while for the node to start revenue contribution. Liberty Times suggests that based on the revenue ramp of the 3nm node, industry analysts expect 2nm to start contributing to TSMC’s revenue as early as Q3 2026.

Though TSMC previously noted that while early 2nm production will dilute gross margins, the higher structural profitability, coupled with improved 3nm margins, should maintain the company’s overall average gross margin. According to TechNews, roughly 30% of TSMC’s capacity for 2nm and more advanced nodes will be sourced from its Arizona fabs.

Meanwhile, TSMC is set to roll out more members of its 2nm family in 2026, with the next-generation N2P (performance-enhanced) node and the A16 (1.6-angstrom) process — which will introduce backside power delivery for the first time — expected to be making debuts in the second half of this year.

According to Tom’s Hardware, in terms of technical gains, N2 is expected to offer a 10%–15% boost in performance at the same power, or a 25%–30% cut in power at the same performance, along with roughly 15% higher transistor density versus N3E for mixed logic, analog, and SRAM designs. For logic-only chips, density gains could reach as much as 20% over N3E, the report adds.

2nm Fever: Flagship Chips Drive Sky-High Prices

As noted by TechNews, TSMC traditionally ramps new process nodes first for mobile devices and small consumer chips, since their smaller die sizes make yields easier to manage. The N2 node, however, breaks with this convention: production is scaling simultaneously for smartphones as well as larger AI and high-performance computing (HPC) designs, reflecting the strong market demand across multiple segments, the report adds.

While TSMC’s 2nm node is expanding its reach, the market’s eyes are squarely on the soaring prices of flagship smartphone processors. According to AndroidHeadlines, Qualcomm’s Snapdragon 8 Elite Gen 6 Pro, which could be the company’s first chip fabricated on TSMC’s 2nm, is reportedly set above $300.

The report suggests that its steep price reflects its cutting-edge 2nm manufacturing, with each wafer reportedly costing around $30,000—meaning the processor alone consumes nearly a third of a premium phone’s production budget.

On the other hand, MacRumors reported that Apple’s A20 flagship chips, built on TSMC’s 2nm process, are expected to cost around $280 each—making them the priciest component in the iPhone. Wccftech suggests MediaTek’s Dimensity 9600, expected in 2026 as well, will also be built on TSMC’s 2nm node.

Read more

• [News] TSMC Reportedly to Raise Sub-3nm Prices 3-10% in 2026, Plans Hikes Through 2029
• [News] TSMC May Fast-Track 1.4nm at Central Taiwan Science Park as 2nm Yields Reportedly Beat Expectations

(Photo credit: TSMC)