Nomura Initiates Coverage on Piotech With Buy, Bullish on China's Top Four Semiconductor Equipment Makers
nashnova research
Nomura initiates coverage on Piotech and resumes coverage on AMEC, ACM Research, and NAURA — all rated Buy, with Piotech ranked first. This means → Nomura is betting not on capacity expansion alone, but on the structural uplift equipment makers get from process upgrades plus domestic substitution.
How big is this market?
Nomura projects China's wafer-fab equipment market will grow from roughly $43.6 billion in 2025 to about $103.5 billion by 2030 — a ~19% CAGR.
Cumulative spending from 2026 to 2030: approximately $407.3 billion. In plain terms = five years of spending that exceeds the prior decade.
The base case assumes mainland China cannot obtain EUV lithography tools from 2026 through 2030. This means → the entire model is built on the strictest export-control scenario; the numbers are already conservative.
Without EUV, why does equipment demand actually rise?
Advanced nodes without EUV must rely on DUV multi-patterning — splitting one lithography step into multiple repeated passes — which multiplies etch, deposition, and cleaning steps.
Nomura's example: a 7 nm-class chip requires ~140 etch steps versus ~40 at 28 nm. Each additional 10,000 wafers-per-month of advanced capacity needs about 60 etch tools; a mature line needs roughly 25.
This reflects a key dynamic: the tighter the EUV restriction, the more process steps each wafer consumes — and equipment makers' value per wafer actually goes up.
How does the memory upgrade add another push?
3D NAND — flash memory that stacks storage cells vertically — directly raises equipment intensity as layer counts climb: etch tools' share of related capex rose from about 35% at 32 layers to roughly 48% at 128+ layers.
In DRAM, storage-capacitor fabrication demands high-aspect-ratio etching (cutting deep into extremely narrow holes) and atomic-layer deposition (ALD — a precision process that coats one atomic layer at a time). Etch and deposition together typically account for over half of a DRAM line's equipment capex.
This means → every generation upgrade at a memory fab ratchets up order density for equipment suppliers — exactly why Nomura weights memory-client exposure so heavily.
Why does Nomura favor Piotech and AMEC most?
Preference ranking: Piotech > AMEC > ACM Research > NAURA — the key criterion is exposure to memory customers.
YMTC and CXMT are the two main drivers of domestic memory expansion, and both are major clients of Piotech and AMEC. Piotech's PECVD business (plasma-enhanced chemical vapor deposition — depositing thin films on wafers using plasma) and AMEC's etch business are directly tied to memory process upgrades.
In plain terms = whoever sits closest to the memory upgrade cycle has the most visible order book — and that is precisely where Piotech and AMEC stand.
What risks should investors watch?
Nomura flags explicitly that some high-aspect-ratio etch and advanced ALD capabilities mentioned in the report are still in development or qualification — not yet in volume production.
This means → within the "domestic substitution" narrative, there is still a gap between R&D progress and mass-production delivery. Investors need to distinguish "entered R&D" from "entered the fab."
The entire model rests on the no-EUV assumption. If export controls ease, the equipment-demand structure could shift, and the logic behind the current ranking would need revisiting.
市场有风险,内容仅供研究参考,不构成投资建议。
