Bernstein Research Report Analysis: Global Semiconductor Equipment WFE Grows 75% in Two Years, DRAM and Logic Chips Are the Biggest Drivers

marsbitPublié le 2026-08-13Dernière mise à jour le 2026-08-13

Résumé

Bernstein significantly raises its global Wafer Front-End (WFE) equipment spending forecast: $148B in 2026 (from $141B), $204B in 2027 (from $175B), and $259B in 2028 (from $198B). This represents a cumulative 75% growth over two years, confirming a strong equipment upcycle. Memory, especially DRAM, is the primary driver, with DRAM WFE expected to surge 53% in 2027 and 37% in 2028. Logic/Foundry follows, driven by advanced node investments. The forecast upgrades are broad-based, but the largest incremental contribution comes from memory spending outside China and logic/foundry expansion within China. Bernstein expects China's WFE demand to be resilient, jumping to $101B in 2028 from $573B in 2027 due to comprehensive expansion across memory, advanced logic, and mature nodes. A key factor is the DUV lithography shortage, which is expected to ease in H2 2026, allowing imports to recover. Top equipment picks include ASML (preferred in Europe) for its exposure to DRAM and advanced logic, and Applied Materials (top US pick) for its broad exposure and attractive valuation. In China, NAURA is favored due to its comprehensive product portfolio and significant advanced logic exposure. Bernstein views the recent sector pullback as an attractive entry point, arguing that market expectations for 2027/2028 demand strength remain too low and that supplier guidance is conservative.

Author: Rita

Bernstein has significantly raised its forecast for global wafer fab equipment (WFE) spending: 2026 raised from $141B to $148B, 2027 from $175B to $204B, and 2028 from $198B to $259B. A cumulative two-year growth of 75%, further confirming the judgment of an equipment spending upcycle. WFE refers to wafer front-end equipment, the core equipment like lithography, etching, and deposition required for front-end chip manufacturing processes. "Logic chips" in the title refer to computing chips like CPUs, GPUs, ASICs, and their wafer foundry capacity.

The upward revision covers nearly all application areas and regions, but the increment comes more from memory chips in regions outside China and from China's foundry and logic chips. DRAM and NAND are the strongest-growing segments, followed by logic and foundry. Bernstein believes the current equipment spending upcycle is sustainable, and the recent pullback in the semiconductor equipment sector offers an attractive entry opportunity.

WFE Grows 75% in Two Years, DRAM Is the Biggest Driver

Memory chips are the core driver of this cycle. Bernstein forecasts DRAM WFE growth of 53% and 37% in 2027 and 2028, and NAND growth of 58% and 42% for the same period. The upward revision in DRAM capital expenditure comes from both a pull from regions outside China and Chinese DRAM manufacturers, while NAND benefits from technology upgrades and capacity expansion.

Logic and foundry growth is slightly lower, with projected increases of 20% and 19% for 2027 and 2028 respectively, mainly driven by advanced node investments. Intel's IDM spending is slightly revised upward, while the China portion remains stable.

DRAM WFE is projected to reach $69B in 2027 (previous forecast $57B), NAND $20B (previous $18B), and logic/foundry $104B (previous $89B). The growth rates for DRAM and NAND WFE are higher than for logic/foundry, which is uncommon in past cycles.

Regions Outside China Contribute the Largest Increment

The main increment comes from regions outside China. Of the $22.8B upward revision for 2027, DRAM accounts for $10.7B and logic/foundry for $9.8B; of the $36.8B upward revision for 2028, DRAM accounts for $18B and logic/foundry for $15B. The $7B upward revision for 2026 is slower-paced, with foundry contributing $3.4B, DRAM $2.6B, and the remainder from other segments.

Guidance from major equipment suppliers is largely consistent with Bernstein's forecast. Tokyo Electron expects WFE of about $120B in 2026, over $150B in 2027, and over $190B in 2028; Lam Research expects close to $110B in 2026 and entering the low $150B range in 2027; KLA expects the low $120B range in 2026, over $150B in 2027, and over $190B in 2028. These figures differ significantly from Bernstein's forecasts. Bernstein believes equipment company guidance is conservative, and upward revisions are likely as order visibility improves in upcoming earnings seasons.

China's WFE Resilience Exceeds Market Expectations

Bernstein forecasts China's WFE demand for 2026-2028 at $57B, $57.3B, and $101B respectively. The figures for 2026 and 2027 are nearly flat, mainly due to DUV lithography tool shortages hampering imports. In 2028, it jumps to $101B with capacity expansion across all segments.

The 2026 forecast is revised down by $1.1B from the previous one, mainly due to weak imports from DUV tool shortages, but domestic equipment makers are expected to achieve 41% year-on-year growth, reaching $16B in revenue. Forecasts for 2027-2028 are significantly revised up by $6.3B and $24B respectively, with expansion across all Chinese segments—memory, advanced logic, and mature logic—all expanding capacity.

Memory remains the strongest segment, with year-on-year growth rates of 88%, 81%, and 62% for 2026-2028. The surge in advanced logic primarily stems from Huawei's advanced packaging solutions, which stack multiple chips to enhance performance, consuming more wafers. For mature logic, domestic NAND manufacturers outsourcing CMOS production to foundries, AI-driven new demand, and the industrial and automotive upcycle are collectively pushing mature node supply to tighten, prompting foundries to resume aggressive capacity expansion.

Bernstein expects China's WFE imports to recover in the second half of 2026, with annual imports hitting a new high of $41B. Global equipment suppliers' China revenue in 2026 will likely be flat year-on-year, better than the previous expectation of a "year-on-year decline," and may be further revised upward in coming quarters.

ASML, Applied Materials, and Naura Each Have Their Positioning

ASML is Bernstein's top pick in European semiconductor equipment, with projected revenue CAGR of 34% from 2025 to 2028, supported by rising DRAM and advanced logic capital expenditure and increased lithography intensity. Both EUV and DUV will see strong growth, with EUV shipments nearly doubling from 48 units in 2025 to 118 units in 2028. While China revenue share is expected to decline to 20% per company guidance (33% in 2025, 41% in 2024), there is significant upside potential from China's continued demand for DUV systems.

Applied Materials sees the largest target price increase, raised from $525 to $675 (+29%). Lam Research is raised from $360 to $385, KLA from $225 to $250, all three maintaining Outperform ratings. Among Bernstein's US equipment stocks, Applied Materials ranks first due to its largest exposure to advanced logic, DRAM, and packaging, coupled with relatively cheaper valuation; Lam Research has stronger execution and NAND upgrade exposure but higher valuation; KLA's growth this year is slightly slower (due to longer lead times), but it has the largest advanced logic exposure and higher upside potential post-2027.

Among Chinese equipment stocks, Bernstein's ranking is Naura优于 Zhongwei优于拓荆科技(Tuojing Technology), all three maintaining Outperform ratings. Naura benefits the most from its broadest product portfolio (covering deposition, dry etch, thermal, cleaning) and largest advanced logic exposure (~40%), making it the biggest beneficiary during accelerated advanced logic capacity expansion. The recent sector pullback provides an attractive entry opportunity, as the market's underestimation of 2027 demand strength will be gradually corrected in the second half of the year.

Bernstein believes the global semiconductor equipment sector is in one of its strongest upcycles since 2008. The intensity of DRAM and advanced logic capital expenditure is systematically underestimated by the market, the resilience of the China market is misunderstood as unsustainable, and major equipment supplier guidance is conservative. Behind the 75% WFE growth in two years is the structural expansion undergoing across the entire semiconductor manufacturing chain.

Disclaimer

This article is Tide Research's organization and interpretation of a third-party brokerage research report (Bernstein, August 11, 2026), combined with compilation of public market information. The ratings, target prices, profit forecasts, and related judgments cited in the article are the views of that brokerage's analyst, representing only the stance of their institution, and do not represent the views of Tide Research, nor constitute any investment advice.

Markets involve risks; decisions should be independent. This article should not serve as the basis for buying or selling any securities.

Questions liées

QWhat is the key driver behind the projected 75% cumulative growth in global Wafer Front End (WFE) spending over the next two years according to Bernstein's report?

AThe key driver is the strong capital expenditure in memory chips, particularly DRAM and NAND. Bernstein's report states that DRAM and NAND are the strongest growth segments, with DRAM being the primary contributor to the WFE growth cycle.

QWhich region is identified as providing the largest incremental contribution to the WFE spending forecast increases for 2027 and 2028?

AThe largest incremental contribution comes from regions outside of China. For instance, in 2028, of the $36.8 billion upward revision, DRAM accounted for $18 billion and logic/foundry for $15 billion from regions excluding China.

QHow does Bernstein's report describe the resilience and outlook for China's WFE demand, especially regarding the significant jump in 2028?

ABernstein states that China's WFE demand shows resilience beyond market expectations. Demand is forecast to jump to $101 billion in 2028 from $57.3 billion in 2027, driven by comprehensive expansion across all segments including memory, advanced logic (spurred by Huawei's advanced packaging), and mature logic.

QAmong the major global equipment suppliers (Tokyo Electron, Lam Research, KLA), how do their provided WFE market guidance figures compare to Bernstein's own predictions?

AThe guidance from major suppliers like Tokyo Electron, Lam Research, and KLA is significantly lower and more conservative than Bernstein's predictions. For example, Tokyo Electron guides for over $190 billion in 2028, whereas Bernstein forecasts $259 billion. Bernstein believes equipment companies are being conservative and will likely raise their guidance later.

QAccording to Bernstein's analysis, which specific company is ranked first among U.S. equipment stocks and what are the reasons given for this preference?

AApplied Materials is ranked first among U.S. equipment stocks. The reasons include its largest exposure to advanced logic, DRAM, and packaging businesses, coupled with a relatively cheaper valuation compared to its peers.

Lectures associées

Le trimestre à 2 milliards de dollars de Coherent : Revenus, marges et investissements en actifs fixes

Le fabricant américain de composants photoniques Coherent a publié ses résultats du quatrième trimestre et de l'exercice 2026, clos le 30 juin. Le chiffre d'affaires trimestriel a atteint 20,46 milliards de dollars, en hausse de 33,8 % sur un an. La croissance s'est inscrite dans une trajectoire ascendante continue tout au long de l'exercice. L'analyse segmentaire révèle que la quasi-totalité de cette croissance nette (+5,16 Md$) provient de la division "Centres de données et Communication", dont le chiffre d'affaires a progressé de 5,97 milliards de dollars. La division "Industriel" a, à l'inverse, enregistré un léger recul. Au T4, "Centres de données et Communication" représentait environ 79 % du chiffre d'affaires total. La rentabilité s'est améliorée plus rapidement que le chiffre d'affaires. La marge brute non GAAP est passée de 38,1 % à 40,2 %, tandis que la marge opérationnelle non GAAP a augmenté de 18,0 % à 21,8 %, bénéficiant également d'une maîtrise des dépenses administratives. Pour soutenir cette croissance, l'entreprise a significativement accru ses investissements. Les ajouts d'immobilisations corporelles ont atteint 1,103 milliard de dollars pour l'exercice, soit environ 2,5 fois le montant de l'année précédente. Cet effort d'investissement s'est accompagné d'une forte augmentation des stocks et d'une réduction substantielle du flux de trésorerie lié aux activités opérationnelles, passant de 634 millions à 80 millions de dollars. Les résultats démontrent une expansion robuste portée par le segment des centres de données, avec une amélioration concomitante de la rentabilité et des investissements massifs en capacités de production. La direction prévoit un chiffre d'affaires de 2,2 à 2,4 milliards de dollars pour le premier trimestre de l'exercice 2027.

marsbitIl y a 2 mins

Le trimestre à 2 milliards de dollars de Coherent : Revenus, marges et investissements en actifs fixes

marsbitIl y a 2 mins

Notes de podcast | Un cadre de Bitwise : Les mauvaises nouvelles ne font plus peur au Bitcoin, les institutions achèteront les grandes cryptos lors du prochain marché haussier, tandis que les capitaux on-chain privilégieront la couche applicative

**Résumé du podcast : Les perspectives de Bitwise sur le marché crypto** Lors de cet enregistrement (août 2026), le bitcoin évolue autour de 65 000 $, en retrait de plus de 50% par rapport aux sommets. Pour Matt Hougan (CIO) et Ryan Rasmussen (responsable recherche) de Bitwise, cette résistance aux mauvaises nouvelles (ventes importantes, piratage, incertitudes réglementaires) est un signe caractéristique de fond de marché. Les vendeurs urgents ont disparu, laissant place aux détenteurs de long terme. Le prochain cycle haussier sera, selon eux, scindé en deux : 1. **Les capitaux institutionnels** (gestionnaires de patrimoine, fonds) se concentreront sur les actifs larges et liquides comme le Bitcoin (considéré comme un "or numérique") et l'Ethereum, via les ETF. Leur processus d'adoption, long (environ 2 ans d'éducation), est désormais mature, promettant des entrées de fonds massives et durables. 2. **Le capital natif de la chaîne (on-chain)** se tournera vers les applications DeFi génératrices de revenus réels (comme Hyperliquid, Uniswap, Aave), attiré par leur modèle économique et leurs flux de trésorerie. Le plus grand catalyseur à venir n'est pas interne à la cryptosphère, mais provient des plateformes de gestion de patrimoine traditionnelles (Morgan Stanley, UBS, etc.) qui gèrent environ 20 000 milliards de dollars. Une allocation même modeste (1-2%) de leurs portefeuilles modèles vers les crypto-monnaies représenterait des centaines de milliards d'afflux. Le cycle de 4 ans se compresse, avec des corrections moins profondes (55% vs 70-80% auparavant) et une volatilité réduite, alimentée par l'entrée d'investisseurs institutionnels au horizon de placement plus long. Le cadre réglementaire (projet Reg Crypto) évolue également positivement. En bref, l'avenir verra une divergence : une montée lente et institutionnelle des grands actifs, et un boom des applications DeFi porté par les acteurs du secteur, dans un contexte macroéconomique de déficits publics soutenant les actifs non souverains.

marsbitIl y a 11 mins

Notes de podcast | Un cadre de Bitwise : Les mauvaises nouvelles ne font plus peur au Bitcoin, les institutions achèteront les grandes cryptos lors du prochain marché haussier, tandis que les capitaux on-chain privilégieront la couche applicative

marsbitIl y a 11 mins

Trading

Spot
活动图片