Solar Wafers

What it is

A solar wafer (光伏硅片) is a thin slice of crystalline silicon on which a solar cell is built. Polysilicon is melted and pulled into single-crystal ingots, which are cut into wafers (Daqo 2026 interim report, p. 9). Most have historically been Czochralski (Cz) silicon, named after the growth process. The quartz crucible adds oxygen to the ingot, which degrades boron-doped (p-type) wafers but not phosphorus-doped n-type ones (PVEducation, Honsberg and Bowden). The industry has moved to n-type: n-type wafers rose from 30% of output in 2023 to 70% in 2024, while sizes grew and rectangular formats spread (IEA PVPS Task 1, October 2025, p. 8).

TCL Zhonghuan reports the two steps separately. In the first half of 2026 monthly output per crystal-growing machine rose 11.8% and slicing machine output 5.9% (TCL Zhonghuan 2026 interim report, p. 11).

Why it matters for the chain

The wafer is the base of the solar cell, so its specification follows cell technology: TCL Zhonghuan says it launched wafers tailored to TOPCon, back-contact (BC) and heterojunction (HJT) cells (p. 12). Integrated makers use most of their own wafers. Of the 48.91 GW of wafers LONGi shipped in the first half of 2026, 18.98 GW went to other companies (LONGi 2026 interim report, p. 12).

Where China stands

Global wafer output was 804 GW in 2024, and China made 96.5% of it. Most of the rest came from Vietnam, Thailand and Malaysia, where major solar makers have plants (IEA PVPS p. 8).

• TCL Zhonghuan (TCL中环, 002129.SZ) shipped 53.9 GW in the first half of 2026 and says it has the largest market share. Its overseas wafer shipments rose 3.7 times, with new customers in India and Turkey (p. 12).
• LONGi (隆基绿能, 601012.SH) says it has shipped more monocrystalline wafers than anyone over the past ten years (LONGi p. 11).

LONGi has stopped expanding: it has paused investment in a 100 GW wafer slicing project in Xixian and a 20 GW ingot project in Xi'an; the cell plants agreed alongside them are partly in production (LONGi p. 38).

Economics

Tongwei, citing CPIA, says wafer and cell prices in July 2026 were nearly 30% below January's, while polysilicon fell more than 40% (Tongwei 2026 interim report, p. 9). Cheaper polysilicon did not rescue wafer margins:

• TCL Zhonghuan's wafer revenue fell 12.62% to ¥5.05bn at a gross margin of -35.54%, down 11.80 points. It spent about ¥1.36 on cost of sales for every ¥1 of wafer revenue (p. 16; our calculation).
• Its modules came close to breaking even, at a gross margin of -1.05% (p. 16).
• LONGi does not report wafer margins separately. Its solar products as a whole earned a gross margin of about -1% (our calculation from LONGi p. 176).

What to watch

• Polysilicon prices. Eight polysilicon makers pledged in August not to sell below full cost (Daqo p. 10). Our view: wafer makers already selling at a gross loss cannot absorb a higher input price, so a floor upstream only holds if wafer, cell and module prices rise with it.
• Moving downstream. In the second quarter TCL Zhonghuan's module revenue overtook its revenue from solar materials (TCL Zhonghuan pp. 11, 12). It completed the takeover of Yidao New Energy in early July and is converting all its cell capacity and half its module capacity to BC (pp. 12, 13).
• Overseas cell making. China's cell exports rose 36.8% in the first half of 2026 while module exports fell 2.5%, as other regions build their own supply chains (LONGi p. 10). TCL Zhonghuan's wafer growth abroad came from customers in India and Turkey (p. 12).