China semiconductor equipment: memory capacity expansion drives a significant upward revision to WFE forecasts
AI summary card
China semiconductor equipment: memory capacity expansion drives a significant upward revision to WFE forecasts
Bernstein believes China WFE is entering a new structural growth phase driven jointly by AI, the memory supercycle, and domestic substitution, and reiterates AMEC, Piotech and NAURA as Outperform.
- China WFE spending forecast raised to USD 58/67/77bn for 2026E/2027E/2028E, materially above the prior forecast.
- CXMT and YMTC are expected to add one fab each in 2026 and two more each in 2027, with the higher memory capex mainly flowing through to 2027/2028 WFE demand.
- Revenue of domestic equipment vendors is expected to rise from USD 15bn in 2026E to USD 22bn in 2027E and USD 33bn in 2028E, while self-sufficiency is expected to increase from 26% in 2026E to 43% in 2028E.
- Domestic substitution is being driven by U.S. export controls, co-development between fabs and domestic equipment vendors, government subsidies, and Big Fund support.
- The report warns that domestic equipment stocks may pull back in the near term after a strong rally, but order growth remains stronger than expected and there is still upside over the coming months.
Report interpretation
Overview
This report updates Bernstein's China wafer fabrication equipment (WFE) model. The main change is an upward revision to China's memory capex and its 2027/2028 China WFE demand forecast based on channel checks over the past quarter. The report argues that strong DRAM/NAND demand, AI-related compute demand, improved financing and profitability among domestic memory makers, and U.S. supply-chain restrictions will together push CXMT, YMTC and other Chinese memory IDMs to accelerate capacity expansion. Domestic equipment vendors are therefore likely to continue benefiting from orders, revenue and localization rates.
Core views
The report's core views are: first, China WFE is entering a structural growth phase driven by the overlap of the AI cycle, the memory supercycle and domestic substitution; second, China WFE demand for 2026E/2027E/2028E is expected to reach USD 58/67/77bn, materially above the prior USD 55/59/61bn forecast; third, global equipment vendors' China revenue is expected to be broadly flat, while domestic equipment vendors' revenue will grow rapidly, pushing self-sufficiency to 43% in 2028E; fourth, domestic substitution in DRAM may see a step-up in 2027; fifth, AI chip localization will increase domestic HBM, eSSD, DRAM and NAND demand, creating further upside risk to memory capex.
Analysis framework
The report uses a top-down forecast for China WFE demand, capex breakdowns by application, import-versus-domestic supply decomposition, and bottom-up estimates of revenue for domestic and global equipment vendors. On the demand side, it focuses on advanced logic, mature logic, DRAM, NAND and non-China customers; on the supply side, it compares China semiconductor equipment revenue for the top five global vendors with that of major listed Chinese equipment vendors.
Methodology notes
Forecast annual China WFE total demand, import demand and domestic supply revenue.
The model splits China WFE demand into revenue for imported tools from global suppliers and revenue for domestic equipment vendors, and further incorporates downstream expansion cycles in memory, advanced logic and mature logic.
Validate domestic IDM capacity expansion and equipment order trends through channel research.
The report says channel checks over the past quarter showed Chinese memory IDMs clearly accelerating expansion plans for the coming years, while domestic equipment vendors' order guidance also became more constructive.
Estimate the revenue of domestic and global equipment vendors in China's semiconductor equipment market from the bottom up.
The domestic sample includes AMEC, NAURA, Piotech, ACMR Shanghai, Kingsemi, PNC, Hwatsing Tech, Leadmicro Nano, Skyverse and Wuhan Jingce; the global sample includes ASML, Lam Research, Applied Materials, Tokyo Electron and KLA.
Measure localization progress by the share of domestic equipment vendor revenue in total China WFE demand.
The report expects China semiconductor equipment self-sufficiency to rise from about 21% in 2025 to 26% in 2026E, 34% in 2027E and 43% in 2028E.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- NAURAA domestic WFE leader; the report reiterates Outperform with a target price of CNY 680.00.
- Strengths
- The broadest product portfolio, covering PVD, CVD, ICP dry etch, thermal processing and cleaning, with a diversified customer base across logic, DRAM and NAND.
- Weaknesses
- The report does not identify company-specific shortcomings, but at the industry level some equipment still lags global leaders in yield, stability and uptime.
- Comparison
- Compared with most domestic equipment vendors, NAURA covers a broader set of equipment categories and is therefore better positioned to benefit systematically from China's WFE localization.
- Risks
- Domestic equipment stocks may face near-term pullbacks after a strong rally; the pace of localization, client expansion schedules and equipment performance validation remain key uncertainties.
- AMECA core domestic WFE name; the report reiterates Outperform with a target price of CNY 500.00.
- Strengths
- Primarily focused on CCP and ICP dry etch, while rapidly expanding into deposition tools such as ALD, LPCVD and EPI; regarded as a Chinese WFE company with strong technical capability and global recognition.
- Weaknesses
- Its product line is more concentrated than NAURA's, so it depends more heavily on progress in etch and new deposition businesses.
- Comparison
- It stands out in technical reputation and is well positioned to gain share in high-end etch and deposition localization.
- Risks
- Advanced-node equipment validation, customer adoption pace and valuation volatility may affect returns.
- PiotechA domestic WFE growth name; the report reiterates Outperform with a target price of CNY 580.00.
- Strengths
- Focused on deposition tools such as PECVD, HDPCVD, SACVD and ALD, and also expanding into W2W and C2W hybrid bonding equipment, with a strong track record of product innovation.
- Weaknesses
- Its scale and product breadth are still expanding relative to NAURA and AMEC.
- Comparison
- More growth- and innovation-driven, benefiting from the localization of deposition equipment and expansion into advanced packaging tools.
- Risks
- Commercialization of new products, realization of advanced packaging demand and customer validation cycles are the main uncertainties.
- Tokyo ElectronA global SPE supplier; the report rates it Outperform with a target price of ¥59,200.
- Strengths
- The world's fourth-largest SPE supplier and the largest Japanese SPE supplier, with meaningful exposure across the six major product categories.
- Weaknesses
- Accelerating local substitution in China may pressure some long-term share.
- Comparison
- Versus U.S. suppliers, a weaker yen and competitive pricing may help share and margins.
- Risks
- Normalization of China revenue, the global WFE cycle and FX changes may affect performance.
- KokusaiA global equipment name; the report rates it Outperform with a target price of ¥8,240.
- Strengths
- Batch ALD adoption is expected to rise, especially in GAA at advanced nodes, and NAND is also an important application for Batch ALD.
- Weaknesses
- Sensitive to the recovery in NAND capex and the pace of adoption at advanced nodes.
- Comparison
- It has an advantage in the ALD niche and benefits from NAND recovery and process changes at advanced nodes.
- Risks
- A weaker-than-expected NAND recovery or slower customer expansion.
- ScreenA global cleaning equipment name; the report rates it Market-Perform with a target price of ¥12,600.
- Strengths
- Potential upside comes from panel-level packaging.
- Weaknesses
- Cleaning intensity has not improved materially, and competition is intense.
- Comparison
- It faces significant competition from TEL, Lam, and global and Chinese competitors such as ACMR and NAURA.
- Risks
- Competition in cleaning tools, pricing pressure and slower-than-expected panel-level packaging adoption.
- AMATA global semiconductor equipment name; the report rates it Outperform with a target price of $525.00.
- Strengths
- Benefits from long-term WFE growth, SAM expansion, a stronger services narrative and capital returns.
- Weaknesses
- Local substitution in China and export restrictions may limit some China-related opportunities.
- Comparison
- As one of the global leaders, it benefits from global WFE expansion, but its China share may face local substitution.
- Risks
- Export controls, the global capex cycle and changes in China exposure.
- LRCXA global semiconductor equipment name; the report rates it Outperform with a target price of $340.00.
- Strengths
- CY25 commentary has been fairly supportive, and an NAND upgrade cycle may be starting.
- Weaknesses
- Sensitive to NAND capex and the China restriction environment.
- Comparison
- It has leverage to the NAND-related equipment cycle, but also faces domestic substitution and cyclical volatility.
- Risks
- A NAND recovery below expectations, export controls and customer capex volatility.
- ASMLThe global lithography leader; the report rates it Outperform with a target price of EUR 1,700.00.
- Strengths
- Clear leadership in EUV technology, with lithography playing a critical role in advanced nodes.
- Weaknesses
- The report takes a relatively cautious stance on ASML growth; its 2030 revenue forecast sits at the low end of company guidance and about 20% below market consensus.
- Comparison
- It has the strongest technology moat, but the China and EUV assumptions are conservative, so growth expectations are below consensus.
- Risks
- EUV demand, China revenue, export restrictions and the risk of missing long-term guidance.
Key data
- China WFE demand forecast2026E/2027E/2028E: USD 58/67/77bnSignificantly raised from the prior USD 55/59/61bn forecast, mainly driven by stronger memory capex.
- Forecast revision magnitude2026E +USD 2.9bn; 2027E +USD 7.3bn; 2028E +USD 16.4bnThe new 2028E forecast is about 26% higher than the old forecast.
- China WFE market in 2025About USD 50bnOf which imports were about USD 39.2bn, up 3% YoY; domestic revenue was about USD 11bn, up 49% YoY.
- 2026E domestic equipment vendor revenueAbout USD 15bn, up roughly 36%The report believes visibility is relatively high because 2026 revenue is largely derived from 2025 orders.
- 2027E/2028E domestic equipment vendor revenueAbout USD 22/33bnCorresponding to about 47% and 50% YoY growth, respectively.
- China WFE self-sufficiency rate2026E 26%; 2027E 34%; 2028E 43%Localization is driven by post-sanctions co-development, technology progress and policy funding support.
- China share of global WFE demandAround 38%-39% in 2026E-2028EGlobal WFE growth is faster, but China's absolute scale continues to rise.
- Memory WFE demandApproximately USD 14.4/24.3/34.0bn in 2026E/2027E/2028EMemory's share of China WFE is expected to rise from 25% in 2026E to 44% in 2028E.
- Domestic memory expansion assumptionsCXMT and YMTC are expected to add one fab each in 2026 and two more each in 2027Expansion may continue in 2028, but visibility on scale is lower.
Impact & implications
For investors, the report positions China's domestic semiconductor equipment chain as the main beneficiary of China's WFE growth over the next several years. If memory expansion and localization proceed as expected, domestic equipment vendors' order and revenue growth should materially outpace that of global equipment vendors. After a strong rally, share prices may see a near-term pullback, but the report believes order growth remains stronger than the market expects and that AMEC, Piotech and NAURA still have further upside.
Risks
- Domestic equipment stocks in China may pull back in the short term after a strong rally.
- Visibility on 2027E/2028E import demand is limited, and global equipment vendors' China revenue may differ from assumptions.
- Domestic equipment may still lag global leaders in yield, stability and uptime.
- If the memory supercycle, AI demand or domestic HBM/eSSD demand are weaker than expected, the logic for raising memory capex will be affected.
- The scale and pace of CXMT, YMTC and other expansion plans remain uncertain, especially beyond 2028.
- Export controls, supply-chain restrictions, financing conditions and changes in policy subsidies may alter the expansion and localization pace.
- Some import data may be distorted by front-loaded inventory buildup, leading to bias in the assessment of true demand.
What to watch
- Construction of new fabs, equipment orders and ramp progress at CXMT and YMTC.
- Order guidance, revenue recognition and customer validation progress for domestic equipment vendors such as AMEC, NAURA and Piotech.
- DRAM and NAND prices, profitability and capex cycles.
- Changes in demand for HBM, eSSD, DRAM and NAND driven by AI chip localization.
- Whether China semiconductor equipment self-sufficiency can follow the path of 26% in 2026E, 34% in 2027E and 43% in 2028E.
- The latest guidance from global equipment vendors on China WFE demand and China revenue mix.
- Changes in U.S. export controls, government subsidies, Big Fund support and domestic fab procurement policies.