Goldman Sachs: Upgrades WFE Forecast, Sees Further Upside in Semiconductor Equipment Stocks

Overview

Goldman Sachs has published an in-depth report on the semiconductor equipment sector, concluding that despite significant year-to-date gains, equipment stocks still offer upside potential driven by sustained AI-related capital spending and robust memory/foundry demand. The report substantially raises its 2026–2028 global wafer front-end equipment (WFE) spending forecast and employs a scenario-based framework linking 'hyperscaler capex → WFE spending → company earnings → valuation' to quantify risk-reward profiles across names. Results show that Applied Materials (AMAT) and Tokyo Electron offer the most asymmetric upside opportunities, while Chinese vendors AMEC and NAURA remain attractively valued due to rising domestic content and unique growth drivers.

Core views

Global WFE spending forecasts have been significantly revised upward. Based on memory makers’ expansion announcements and positive feedback from equipment suppliers, Goldman Sachs now forecasts global WFE spending at $141B/$186B/$208B for 2026/2027/2028 (up +28%/+32%/+12% YoY), versus prior estimates of $132B/$160B/$174B. Growth is primarily driven by DRAM, NAND, and advanced logic foundry, with DRAM seeing the largest forecast upgrade due to accelerated HBM and conventional DRAM capacity expansions; foundry WFE upgrades benefit from TSMC’s accelerated cleanroom build-outs for N2/N3 nodes. U.S.-listed equipment stocks show divergent risk-reward profiles. Multi-scenario valuation analysis suggests AMAT’s current share price embeds below-base-case capex assumptions. Even in a bear case where hyperscaler capex growth slows sharply to +10%/+5%, AMAT still offers ~29% upside; in a bull case with an extended AI cycle, upside reaches 51%. Tokyo Electron also presents an attractive risk-reward profile. In contrast, while KLA maintains dominant market share in process control, its current valuation largely prices in base-case cloud spending acceleration, leaving it vulnerable to ~10% downside in a bear case—hence the Neutral rating. Lam Research exhibits a more balanced risk-reward profile. Chinese equipment vendors benefit from localization and structural growth. For China, the report builds a sensitivity analysis incorporating both capex growth and domestic supply share. In the base case, China’s semiconductor capex grows modestly through 2026–2028, but domestic equipment vendor share rises from 32% to 40%. Even in a zero-capex-growth bear case, AMEC and NAURA can still deliver earnings growth via share gains, implying 11% and 5% upside respectively; in a bull case, upside reaches 53% and 46%. The report notes the market has not fully priced in accelerating capex from AI and memory demand in 2027/28, advanced node expansions, or higher localization rates driven by supply chain security. A potential structural upside comes from Terafab. Elon Musk’s announced Terafab project aims to vertically integrate semiconductor manufacturing to achieve 1TW of AI compute capacity and has already begun discussions with key equipment vendors including Lam, AMAT, and Tokyo Electron. The report views this as a potential structural upside driver for WFE demand beyond 2027—though not included in base forecasts, it represents a meaningful call option.

Analysis framework

The report employs a rigorous three-step 'macro-to-micro' analytical framework. First, it maps total hyperscaler capex to global WFE spending, separating AI-driven incremental demand from traditional core demand to derive WFE totals under different capex assumptions. Second, using historical regression-derived beta coefficients, it translates WFE spending into revenue and EPS forecasts for each equipment vendor, quantifying differences in cyclicality sensitivity (e.g., AMAT/LRCX are more sensitive to WFE cycles). Finally, it applies a 3-year median P/E multiple across scenarios to map fundamentals directly to implied price changes, enabling objective comparison of margin of safety and elasticity across names under varying market conditions. For China, it further introduces 'domestic supply share' as a key variable to capture alpha from localization independent of industry-wide trends.

Methodology notes

• Valuation MethodPE/PEG valuation

  Scenario-based valuation using forward-year (e.g., CY28E) EPS forecasts multiplied by historical median P/E multiples

  Rather than assigning a single static target price, the report combines long-term earnings forecasts with historical valuation medians to calculate implied share prices under bull/bear/base scenarios. This helps investors understand what future expectations are currently priced in and the potential profit/loss magnitude under different outcomes.

• Industry/Value Chain FrameworkUpstream-to-Downstream Transmission

  Chain analysis from end-market hyperscaler capex → WFE spending → equipment vendor earnings

  Semiconductor equipment is a classic derived-demand industry. Instead of viewing equipment orders in isolation, the report traces demand back to the ultimate payers—hyperscalers—and builds a quantitative transmission model: 'Cloud Capex → WFE → Equipment Vendor EPS.' This methodology is central to identifying cycle inflection points and sustainability.

• Industry/Value Chain FrameworkVolume-price decomposition

  Decomposing Chinese equipment vendor revenue into 'China Semiconductor Capex × Domestic Vendor Share'

  In China, vendor growth depends not only on industry cyclicality (volume) but also on localization rate (price/share). The report decouples these two variables via sensitivity analysis, revealing that even if industry capex stagnates, rising domestic share alone can support earnings growth for leading local players.

• Quantitative/Factor/Portfolio TheoryBeta/alpha analysis

  Using beta coefficients to quantify revenue elasticity of each vendor to total WFE spending

  Different vendors exhibit varying sensitivity to industry cycles. Historical regressions yield WFE beta values (e.g., AMAT ~1.2x, KLAC ~0.95x), meaning AMAT’s revenue is more elastic to the same industry growth. This explains why high-beta names are preferred when bullish on cycle continuation, while low-beta names offer relative downside protection.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Applied Materials (AMAT)

  Top pick with best risk-reward; current price implies pessimistic expectations

  Strengths

  Large exposure to deposition and etch; benefits from GAA and stacked memory architectures; integrated solutions account for ~30% of systems revenue

  Weaknesses

  Potential revenue impact from U.S. export restrictions on China

  Comparison

  Offers greater downside protection than KLA in bear case; lower valuation positioning than LRCX

  Risks

  Tighter export controls; rising share of Chinese domestic suppliers

• Tokyo Electron (8035.T)

  Buy-rated; attractive risk-reward even in bear case

  Strengths

  Broad product portfolio; monopolistic position in track/coat/develop; benefits from global advanced node expansions

  Weaknesses

  JPY currency fluctuations affect USD-denominated competitiveness

  Comparison

  Joint top pick with AMAT; comparable valuation elasticity

  Risks

  Prolonged inventory correction in semiconductors; rising rates compress valuation multiples

• KLA Corp (KLAC)

  Neutral-rated; valuation reflects base case, with more downside than upside

  Strengths

  Dominant ~6.5x market share lead in process control; deep moat

  Weaknesses

  Process control near-term growth unlikely to outpace WFE average; current valuation appears rich

  Comparison

  Weaker risk-reward than AMAT and LRCX; potential ~10% drawdown in bear case

  Risks

  Foundry/logic spending misses expectations; changes in litho intensity alter tool mix

• NAURA (002371.SZ)

  Buy-rated; China’s platform-type equipment leader, benefiting from rising localization

  Strengths

  Covers thermal processing, deposition, etch, cleaning, ion implantation; accelerating R&D validation at advanced nodes

  Weaknesses

  High R&D spend pressures near-term margins

  Comparison

  Core Chinese equipment asset alongside AMEC; broader product portfolio but slightly slower in point technology breakthroughs

  Risks

  U.S. export controls expand to mature nodes; customer capex execution lags

• AMEC (688012.SS)

  Buy-rated; etch equipment leader upgrading toward advanced nodes

  Strengths

  Etch tools qualified in overseas 5nm fabs; acquisition of Hangzhou Zhonggui fills CMP gap; high exposure to memory customers benefits from expansions

  Weaknesses

  High dependence on memory industry cyclicality

  Comparison

  More globally competitive than NAURA in etch segment, but less platform diversified

  Risks

  Export controls block advanced-node tool supply; slowdown in Chinese fab capex

Key data

• 2026E Global WFE Spending Forecast$141bn+28% YoY; raised from prior $132bn forecast, driven by DRAM and foundry
• 2027E Global WFE Spending Forecast$186bn+32% YoY; significantly raised from prior $160bn forecast
• 2028E Global WFE Spending Forecast$208bn+12% YoY; raised from prior $174bn forecast
• AMAT CY28E Bull Case Upside~51%Based on hyperscaler capex growth assumptions of +40%/+15%
• AMAT CY28E Bear Case Upside~29%Even if hyperscaler capex growth slows to +10%/+5%, significant margin of safety remains
• China WFE Localization Rate (2028E Base Case)40%Expected to rise steadily from 32% in 2026 under base case
• AMEC Bear Case Upside11%Upside persists even with zero growth in China semiconductor capex, thanks to share gains
• NAURA 2027E Target P/E Multiple36.5xDetermined via regression between global SPE sector P/E and forward earnings growth

Impact & implications

The report argues the semiconductor equipment sector is not yet in a bubble, as current valuations for most names do not fully reflect the durability of AI-related capex or the certainty of memory/foundry expansions. For global investors, this implies continued overweighting or adding exposure to high-operating-leverage leaders that remain undervalued even in bear cases (e.g., AMAT, Tokyo Electron), while remaining cautious on names whose valuations already price in base-case expectations (e.g., KLA). For China-focused investors, the investment thesis for domestic equipment vendors has evolved from pure 'cyclical beta' to 'structural growth driven by share gains.' Even under heightened export controls or muted industry sentiment, leading Chinese firms possess cycle-resilient qualities, and current valuations offer an attractive entry point. Additionally, new entrants like Terafab could reshape the medium-term demand curve, creating re-rating opportunities beyond traditional cyclical frameworks.

Risks

• Significant slowdown or reversal in hyperscaler AI capex growth in 2027/28
• Further tightening of U.S. semiconductor export controls on China, extending to mature nodes or specific equipment types
• Longer-than-expected semiconductor inventory correction delaying equipment demand
• Accelerated substitution by Chinese domestic vendors in mature nodes, pressuring foreign vendor shares
• Rising interest rates compressing valuation multiples of high-growth tech stocks
• NAND flash makers pausing technology upgrades, impacting related equipment orders

What to watch

• Hyperscalers’ quarterly capex guidance and AI monetization progress in earnings calls
• Terafab’s negotiation outcomes with equipment suppliers and order placements
• Expansion cadence of memory makers (Samsung, SK Hynix, Micron) for HBM and conventional DRAM/NAND
• TSMC’s N2/N3 cleanroom construction progress and equipment move-in timelines
• Actual execution of China’s semiconductor capex and domestic equipment qualification rates
• Scope and implementation details of next round of U.S. export control policies